Journal of the

CONTENTS

Reflections on 100 years of Rhodora. Janet R. Sullivan

New England Botanical Club

A review of the =<ee —— of Hackelia venusta ieceeurese Ri EG

chy J. Harrod, Lauri A. Malmquist, and Robert L. Carr ............ 16 Myriophyllum — (Haloragaceae), a rare lowland watermilfoil olivia. Garrett E. Crow and Nur P. Ritter 28 Distribution and — of submerged aquatic vegetation beds in a Con- necticut harbor. Todd A. Randall, John K. Carlson, and Matthew E. roczka 40 The distribution of the bryophytes and vascular plants within Little Dollar ake Apt Mackinac County, Michigan. C. Eric Hellquist and Garrett E. Cro 46 NOTE More molecular evidence for interspecific relationships in Liguidambar (Hamamelidaceae). Jianhua Li and Michael J. Donoghue .......... 87 BOOK REVIEW The Savage Garden 92 NEBC MEETING NEWS 95 Information for Contributors 101 NEBC Membership Form 103

NEBC Offficers and Council Members

MISSOURI BOTANICAL

MAY 0 4 1999

GARDEN LIBRARY

Vol. 101 Winter, 1999 Issued: April 19, 1999

inside back cover

No. 905

The New England Botanical Club, Inc.

22 Divinity Avenue, Cambridge, Massachusetts 02138

RHODORA

JANET R. SULLIVAN, Editor-in-Chief Department of Plant Biology, University of New Hampshire, Durham, NH 03824 ANTOINETTE P. HARTGERINK, Managing Editor Department of Plant Biology, University of New Hampshire,

Durham, NH 03824 Associate Editors HAROLD G. BROTZMAN STEVEN R. HILL DAVID S. CONANT THOMAS D. LEE GARRETT E. CROW THOMAS MIONE

K. N. GANDHI—Latin diagnoses and nomenclature

RHODORA (ISSN 0035-4902). Published four times a year (January, April, July, and October) by The New England Botanical Club, 810 East 10th St., Lawrence, KS 66044 and printed by Allen Press, Inc., 1041 New Hampshire St., Lawrence, KS 66044-0368. Periodicals postage paid at Lawrence, KS. POSTMASTER: Send address

changes to RHODORA, PO. Box 1897, Lawrence, KS 66044-8897.

RHODORA is a journal of botany devoted primarily to the flora of North America. Monographs or scientific papers concerned with systemat- ics, floristics, ecology, paleobotany, or conservation biology of the flora of North America or floristically related areas will be considered.

ACCREDITED with the International Association for Plant Taxonomy for the purpose of registration of new names of vascular plants (ex- cluding fossils).

SUBSCRIPTIONS: $75 per calendar year, net, postpaid, in funds paya- ble at par in United States currency. Remittances payable to RHO- DORA. Send to RHODORA, P.O. Box 1897, Lawrence, KS 66044- 8897.

MEMBERSHIPS: rege! $35; Family $45; Student $25. Application form printed herein

NEBC WEB SITE: Information about The New England Botanical Club, its history, officers and councillors, herbarium, monthly meetings and special events, annual te student award, and the journal RHO- DORA is available at http://www.herbaria.harvard.edu/nebc/

BACK ISSUES: Questions on availability of back issues should be ad- dressed to Dr. Cathy A. Paris, Department of Botany, University of Vermont, Burlington, VT 05405-0086. E-mail: cparis@ moose.uvm.edu.

ADDRESS CHANGES: In order to receive the next number of RHO- DORA, changes must be received by the business office prior to the first day of January, April, July, or October.

RHODORA, Vol. 101, No. 905, pp. 1-15, 1999 REFLECTIONS ON 100 YEARS OF RHODORA

JANET R. SULLIVAN, Editor-in-Chief

One hundred years ago the New England Botanical Club began publication of its journal, Rhodora (volume 1, number 1 was published in January, 1899). The Editor-in-Chief was Benjamin Lincoln Robinson of Harvard University. He was assisted by As- sociate Editors Frank Shipley Collins, Merritt Lyndon Fernald, and Hollis Webster, and members of the Publication Committee, William Penn Rich and Edward Lothrop Rand. The first issue of the journal comprised 20 pages of text and two plates, and was distributed to approximately 600 subscribers who had paid $1.00 each for the year’s 12 issues.

That first issue of Rhodora opened with an editorial announce- ment (Editorial 1899) outlining the purpose of the journal (“‘... with confidence that it will give new stimulus and render material aid to the study of our local flora.”) and describing the types of articles to be published. The expectations seem modest but inclu- sive and are clearly stated: that .. Special attention will be given to such plants as are newly recognized or imperfectly known within our limits, to the more precise determination of plant ranges, to brief revisions of groups in which specific and varietal limits require further definition, to corrections upon cur- rent manuals and local floras, to altitudinal distribution, plant as- sociations, and ecological problems.” It was the intention of the editors to consider all contributions dealing with the “‘... flow- ering plants, ferns, mosses, and thallophytes ...”’; presumably they intended to include the gymnosperms, as well. The an- nouncement also states that ““A decided preference will be given to articles which embody some newly observed fact, tersely stat- a

oe

The contents of the first issue encompass the angiosperms, fun-

i, and algae, with short papers and notes on rattlesnake plantains (Fernald 1899), Myosotis collina (Williams 1899), Sanicula (Brai- nerd 1899), fringed gentian (Deane 1899), wild lettuce (Robinson 1899), Matricaria discoidea (Manning 1899), algae (Collins 1899a), and fleshy fungi (Webster 1899). It is interesting to note, in these days when we are so conscious of the importance of peer review and the avoidance of conflict of interest, that all four of

1

2 Rhodora [Vol. 101

the editorial staff members were contributors to this first issue! Like a club newsletter might, the issue includes a listing of new NEBC officers, reports of meetings of other botanical clubs in New England, and a request for subscribers to report on the oc- currence of inland populations of halophytes. The final announce- ment in the issue is a sort of “‘coming attractions,” describing articles promised for the early issues of Rhodora and making an appeal for the submission of announcements of local floras in preparation.

During the period of the celebration of the Club’s centennial in 1996 I became intrigued by the events that had transpired so long ago, resulting in a club and a journal I had respected, ad- mired, and enjoyed since starting my formal study of botany in college in the 1970s. The year of the Club’s centennial was also my first year as Editor-in-Chief of Rhodora; by the end of this year I had a new appreciation of what is involved in the produc- tion of a scientific journal. What activities of the Club’s members warranted such an undertaking? Why did the NEBC decide to publish what was then a monthly journal?

INITIATING PUBLICATION OF A BOTANICAL JOURNAL

The New England Botanical Club was established in February of 1896 “for the promotion of social intercourse and the dissem- ination of local and general information among gentlemen inter- ested in the flora of New England” (Minutes of the 3rd Club meeting, February 5, 1896). By December of that year the mem- bers judged that the endeavor had been successful: the ‘‘faithful workers in the cause of botany,”’ having been isolated previously, were meeting monthly, except during the summer, to share con- versation and information about the New England flora as well as about new research and developments of botanical and eco- logical interest worldwide. In addition, the members of the Club ha en true to their goal of providing ‘‘an elaboration of the New England flora,” by accumulating specimens for the Club’s herbarium (now NEBC), and by stimulating interest among the membership to provide local floras and plant lists. At the annual meeting in December, 1896, it was stated that ‘‘A plan for con- solidating the information thus secured will shortly be presented to you.” It was apparently from this interest in the dissemination

1999] Sullivan—Centennial Reflections 3

of information about the flora of New England that the idea of publishing a botanical journal evolved.

By the fall of 1897 (Minutes of the 8th Council meeting, Oc- tober 22, 1897), Dr. B. L. Robinson announced that Mr. E S. Collins “... had his marine algae in such shape as to be ready to submit his list to the members of the Club before incorporating it in the checklist of New England plants.’”’ There were several local floras already in existence when the Club was organized two winters earlier, and the members had begun accumulating plant lists to be incorporated into a checklist of New England plants. In May and June of that year, Robinson’s Publication Committee had requested help from the members in accumulating information for a checklist, particularly with verifying specimens and literature accounts. The effort to put together a checklist not only increased the holdings of the Club’s herbarium, but even- tually led to the publication of Rhodora and the revision of Gray’s Manual of Botany

The announcement that FE S. Collins was close to being able

to submit his list for distribution sparked “. .. an exhaustive ex- amination of the various systems of multiplication of writing .”’ By vote, the Council approved “‘... that Dr. Robinson in-

eubigaie the possibilities of getting pain papers to be submit- ted to the Club, hektographed by an assistant and that Mr. Collins’ list be hektographed for distribution as an experiment provided the rate of cost be satisfactory ...’’ [The hektograph was a pre- decessor of the spirit duplicator. Invented during the 1870s, it used a stiff gelatin pad and an aniline dye ink to produce copies.]

By February of 1898 (Minutes of the 10th Council meeting, February 21, 1898), the members of the Council were seriously discussing the publication of a ‘‘. .. Bulletin to publish botanical results arrived at by members of the Club and to become the organ of Descriptive Botany in New England.” Dr. Robinson was ready to bring the matter before the Club in a formal proposal The motion to appoint a committee to explore the possibility of producing a publication was carried unanimously at the Club meeting on March 4. At the next Council meeting (Minutes of the 12th Council meeting, March 23, 1898) Robinson presented an exhaustive report by the Publication Committee, and it was recommended that two members be appointed as a Financial Committee to investigate what support might be available for a serial publication. At the April meeting of the Club, E. L. Rand

4 Rhodora [Vol. 101

submitted a report by the committee which reviewed the different ways the publication could be supported, and recommended that not less than 400 subscriptions be obtained. The committee felt confident that the journal could be published without likelihood of failure with this number of subscriptions, and that “*... any deficiency of income during the early stages could Veiadiliy be met.’ Members of the Club present at the meeting exhibited a great deal of interest in the suggested publication.

In the Club scrapbook there is an unsigned, handwritten report dated ‘‘abt Mar 1898” which outlines the details of the — recommended by the Publication Committee. It specifies a ‘ monthly issue of 16 pages and cover ... with iepivalieckes: to consist of six plates per year to be assigned free of charge to authors whose articles seem most in need of illustration . . .”’ The report also notes the attributes desired of the first cover page, with *‘.. . the other three pages of cover to be used for advertising when appropriate and dignified advertisements are available.” These advertisements were later to prove controversial, and soon were dropped from the journal.

The original proposal recommended that each author receive 25 free reprints of articles exceeding one page. The cost of the publication, including wrapping and mailing of 500 copies, was estimated to be $550.00 for the first year, although this was crossed out and $600.00 was penciled in above. The cost of an annual subscription was set at $1.00.

In April of 1898, a printed circular was distributed to the mem- bers of the Club announcing that the publication of a monthly journal was being considered, and requesting the assistance of Club members in obtaining subscriptions (Figure 1). The note accompanying the flier suggested that an average of 10 subscrib- ers procured by each resident member (meaning members resid- ing within 25 miles of Boston, where monthly meetings were held), and five by each nonresident member would assure the necessary financial support of the publication.

y the June meeting of the Council (Minutes of the 14th Coun- cil meeting, June 23, 1898), E. L. Rand reported that 450 sub- scriptions had been taken, and Robinson asked that the matter of organization of an editorial board be brought to a vote. The Coun- cil was in favor of immediate action and appointed B. L. Rob- inson as Editor-in-Chief, and F S. Collins, M. L. Fernald, and Hollis Webster as Associate Editors. This editorial board, in con-

a L. Goopa - nt. peg R. psimncles - Vice- President. Emice F. Wituiams, - - - Treasurer

[6661

Epwarp L. Ranp, Cieaciaiine Secretary.

Mew England Botanical Cfub.

THE NEW ENGLAND BOTANICAL CLUB is considering the publication of a monthly journal, to begin January 1, 1899. It is to be an octavo of about sixteen pages each issue, and illustrated by full-page plates. It will deal primarily with the flora of New England, especial attention being given to rare plants, extended ranges of distribution, and newly introduced, as well as newly described, species. Articles have been already promised by many of the fore- most New England e tanists, both professional and amateur, and, while a hen standard will be maintained in the matter of scientific accuracy, nee ae technicality of style will be carefully avoided, so that any person who can use ‘Gray’s Manual’? will be able to read the proposed journal with pleasure ea Interest. Not only the eile plants and ferns, but fleshy fungi and other cryptogams, will receive attention, The price of the journal has been fixed at one dollar per annum

While more than two hundred subscriptions have already been promised in advance, the Club does not feel warranted in proceeding with its plan of publication unless assured of much further support. All persons interested in botany and in the maintenance of such a journal in New England, are earnestly solicited to send at once sub- scriptions for at least one year (which, however, need not be paid before January 15, 1899), to

EDWARD L. RAND, Corresponding Secretary N. E. Botanical Club, 740 Exchange Building,

APRIL 15, 1898. Boston, Mass.

suonoapyoy [eTuusyueDj—urearns

Copy of a flier announcing the New England Botanical Club’s intention to begin publication of a monthly journal

igure 1. and soliciting subscriptions. -

6 Rhodora [Vol. 101

junction with the already established Publication Committee, was charged with the responsibility of making the business arrange- ments necessary.

ce the editorial board was established, the Council discussed the name of the journal; in an informal vote the name ‘‘Rhodora’”’ was chosen unanimously, but it was decided to defer the matter to a vote of the Club. At the October Club meeting (October 7, 1898) E. L. Rand reported to the members present that over 600 subscriptions had been secured, “‘... thereby insuring the finan- cial support deemed necessary for a beginning.’’ The editorial staff had “*... assumed the responsibility of the business of the journal ...”’ and Dr. Robinson was called upon to state what had been done in regard to naming the journal. “In response .. . Dr. Robinson said the name of the journal had been given the most earnest consideration from the time the project was first dis- cussed. He explained the very great advantages resulting from the use of a single name, not only as being more direct and definite in the minds of the botanical public, but as being vastly more convenient for the purposes of citation.’’ He described the infor- mal vote in favor of ‘““Rhodora”’ at the June meeting of the Coun- cil, and there followed “*. . . a good deal of discussion among the members in regard to the name proposed.” There were strong opinions on both sides of the issue, and the matter was deferred for final action to the next meeting of the Club.

At the 27th Club meeting on November 4, 1898, the subject of a name for the journal was again taken up. There was a great deal of discussion, during which several names were proposed, and eventually an informal ballot was taken. Jesse Greenman was appointed by the Chair to collect the ballots and report the results (Figure 2). ““Mr. Rich, seconded by Mr. Kidder, then moved that the Club adopt the name ‘Rhodora, Journal of the New England Botanical Club,’ as the official title of the proposed publication

and the motion was carried with only one or two dissenting votes.’

Thus the journal was underway, after a year of formal discus- sion and preparation. The Club ended its third year on a high and hopeful note; the members felt secure that the new journal would be self-sustaining, at least at the start, and that its publication “... would reach far and wide, not only our non-resident mem- bers ... but also the great botanical world, who knows us not.” In the summary statements closing the meeting, the new journal

1999] Sullivan—Centennial Reflections 3

i 8 : \ Pee live ;

Vk "

f

das gue ow atthe

Figure 2. Tally of votes taken on the name of the new journal (from the Minutes of the 27th Club meeting, November 4, 1898).

was noted as the highlight of the Club’s achievements: ‘‘.. . the Club now has a voice ... It remains for the members to make that voice heard for the best interests of our favorite science and the result we hope will justify the establishment and maintenance of the Club. This is a momentous period for us and the prosperity and perhaps even the existence of the Club will depend upon the faithfulness with which each one contributes to the success of our undertaking. The labor we have undertaken is great, but as mem- bers of the New England Botanical Club, we should be untrue to our aims and ideals if we did not make the effort to attain them.”

These powerful and encouraging words still ring true today; while the existence of the Club may not depend on the publication of the journal, both the existence and quality of the journal still depend primarily upon the contributions and dedication of Club members.

PUBLISHING A BOTANICAL JOURNAL

The work had only just begun. The editorial staff and many other members of the Club spent the first year of the journal’s publication encouraging the submission of manuscripts, soliciting new subscriptions, obtaining advertisements to help defray costs, and working to maintain the monthly publication schedule.

Publication costs were higher than originally anticipated, due to the decision to electrotype each page. In addition, the inclusion of an index had not been part of the original calculations, and

8 Rhodora [Vol. 101

toward the end of the first year it was estimated that the deficit might be as much as $200. In April of 1899 (Minutes of the 18th Council meeting, April 21, 1899) the Council voted to cover the expense of publishing the New England Checklist ($35), includ- ing providing one set of reprints free to each member of the Club. Emile Williams was appointed to obtain subscriptions adequate to meet the journal’s deficit for 1899. Over the next few years Rhodora consistently ran a deficit, and regular appropriations of funds were made by the Club to support the journal, sometimes at the expense of increasing the holdings of the Club’s herbarium. Nevertheless, the subscription rate was not raised until 1912, when the price of receiving the year’s issues went from $1.00 to $1.50 per year.

In those early years, the journal carried advertisements to help defray the costs of publishing plates. Both W. P. Rich and E. L.

and had addressed the Club, describing to the members the de- sirability of securing advertisements in order to enable the editors to increase the number of pages and to publish more and better plates. The earliest advertisements were for booksellers, collec- tions of dried and live plant specimens, and the ‘“‘Nurseryman’s Directory.”’ After the excursion to Mount Katahdin by five NEBC members in July 1900, an advertisement appeared offering ‘‘Ka- tahdin on Horseback” (volume 3, 1901). The charge for a 4” X 3/4” space was quoted as $4.00 per year (volume 5, 1903).

The largest ads, running from one to four full pages, were purchased by the Bangor & Aroostook Railroad and offered rail passage to remote plant collecting sites. These ads immediately proved controversial, because they itemized locations of rare spe- cies in addition to offering safe, comfortable passage. An editorial published in volume 3 (Editorial 1901) defended the ads and the naming of rare species, stating that the plants in question were abundant at the sites to be visited, and adding that some of the species mentioned were considered weeds by farmers or were timber pests. The ads continued to be published during 1901, but the controversy led to diminishing numbers of ads of all kinds over the next few years and advertising was dropped from the journal by volume 9 (1907).

From a scientific standpoint, the Club’s members were ‘entirely satisfied’ with their new journal. The early volumes of Rhodora were filled with short papers and notes outlining the distribution of particular species in the New England states, describing “‘note-

1999] Sullivan—Centennial Reflections 9

worthy specimens” with unusual morphology, and providing ad- ditions to the checklist. One of the most quaint — described the effects of an inadvertent Boletus poisoning at a brunch held at the home of one of the editors, E S. Collins pee 1899b), and argued for better identification manuals for the region’s flora.

The journal also featured short articles detailing meetings, field trips, and histories of other botanical clubs in New England, such as the Connecticut Valley Botanical eae the Josselyn Botan- ical Society, the Vermont Botanical Club, and the Boston My- cological Club. Short book reviews and gaa started ap- pearing in volume 2, as well. Longer articles commemorating the lives of deceased Club members each featured a formal portrait and signature.

Volume 3 devoted a considerable number of pages to a diary kept of the botanical excursion to Mount Katahdin the year before by five members of the Club (Churchill 1901). The editors also allowed publication of seven plates accompanying this article. In addition, members of the party contributed four other articles on the botanical aspects of the trip (Collins 1901; Fernald 1901; Kennedy and Collins 1901; Williams 1901). This was the begin- ning of a tradition of publishing botanical commentaries on Gray Herbarium and New England Botanical Club expeditions.

Originally, the scientific articles published in Rhodora were mostly field observations, though the results of laboratory exper- iments and study of herbarium specimens started appearing in very early issues. Although notes of one or a few paragraphs continued to be a feature of the journal for many years, the length of articles and their scientific content increased significantly over the first few years of publication.

Nevertheless, manuscripts were difficult to come by in those early years. The members of the editorial staff were frequent con- tributors to the early issues; a glance at the index to volume 1 shows that Robinson wrote four articles and notes, Collins wrote five, Fernald wrote 15, and Webster wrote eight. The items con- tributed by the editors amounted to 28% of the total published that year; the percentage was almost as high for volume 2 (25%). This must have been a considerable burden for the editors; B. L. Robinson addressed the problem at a Club meeting, asking the members “... to bear in mind Rhodora which is much in want of copy ...” (Minutes of the 45th Club meeting, October 5, 1900). After a few years the journal had a group of supporters

10 Rhodora [Vol. 101

who submitted articles regularly, and in 1929 it was noted that articles by 399 different botanists had been published in Rhodora (Editorial 1929). By the 1940s the number of pages published per year had increased considerably, as had the number of plates.

By far the most prolific contributor to the journal was Merritt Lyndon Fernald. An apparently tireless researcher, Fernald pub- lished an average of 13 articles and notes per year (range 4—25) from the time the journal was established until his death in 1950. He also served on the editorial board during this entire period, first as an Associate Editor (1899-1928) and then as Editor-in- Chief (1928-1950) after the resignation of B. L. Robinson. His contributions encompass the full spectrum of types of articles published in the journal during its first 52 years; commentaries on botanical expeditions, notes on the distribution of taxa, floristi and taxonomic treatments, descriptions of new species and vari- eties, details of anomalous plant distribution and morphology, and descriptions of new plant collecting techniques all can be found among the papers he wrote for Rhodora. One may wonder wheth- er the journal could have survived the early paucity of manuscript submissions if it had not been for Fernald’s contributions.

In addition to the variation in availability of publishable ma- terial, the journal suffered an inevitable fluctuation in subscrip- tions. As late as 1928 members of the NEBC Council were still struggling to improve circulation. Originally, having a subscrip- tion to Rhodora was not tied to membership in the Club. [It was not until 1968 that women were admitted to the NEBC as mem- bers, and not until 1996 that membership was automatic upon application.] It was expected after the first year of publication that there would be a “considerable falling off in subscriptions, many of the first year’s subscribers finding it altogether above their interest and understanding.” The list of original subscribers surely included some who fell into that category, but the sub- scription list also must have expanded beyond New England fair- ly early. At the end of the first year (volume 1, number 12), W. P. Rich called for a prompt renewal of subscriptions, listing the cost as “‘$1.00 per year for the United States and Canada, $1.25 for other countries.” Manuscripts were contributed by botanists from outside of New England almost from the start; as early as volume 2 a note appeared by William M. Canby of Wilmington, Delaware (Canby 1900), and in volume 3 a note was published by Charles Bessey of the University of Nebraska (Bessey 1901).

1999] Sullivan—Centennial Reflections 11

In volume 21, FE S. Collins published the first article about plants occurring outside of North America (Collins 1919). Back issues were already scarce by volume 3, and a special notice published in 1901 encouraged interested readers to send the $1.50 necessary to secure a copy of the fast-disappearing issues of volume 1. “ early response will be necessary ...’’ warned the editor.

In 1900 (volume 2) the fectenel fell behind in its publication schedule because of a fire at the printing office. “It appears that the entire April issue was destroyed, but as full sets of proofs had been printed and sent the issue will be immediately reprinted with a delay of perhaps three weeks.’’ (Minutes of the 41st Club meet- ing, April 6, 1900). Luckily, the plates had been stored in a vault and were unharmed by the fire. The journal continued to appear on a monthly basis, more or less on schedule, until 1962 when publication was changed to a quarterly schedule. At this time the subscription rate was raised to $6.00 per year.

RHODORA TODAY

In some ways the business of publishing a botanical research journal has not changed in the past 100 years, and in other ways it has changed dramatically. We still suffer fluctuations in suitable manuscript submissions and journal circulation, and we still need to maintain a regular publication schedule to satisfy both our subscribers and the U.S. Postal Service. Beyond that, however, the original editorial staff of the journal probably would be amazed at the changes in the complexities of the process. The papers published in Rhodora today involve more experimental results than pure description, reflecting that trend in botanical research over the past fifty years. The degree of specialization in research necessitates enlisting the help of reviewers beyond the editorial board. The papers published today are longer than those in the early issues of the journal, and typically include tables and figures. Likely, it has been the increased level of detail in man- uscripts, combined with the use of word processors rather than secretaries, that has contributed to the workload and, thus, to the decreasing tenure of editors of Rhodora (Figure 3), despite the redistribution of some of the more tedious tasks among the press and editorial board members. Certainly, the authors and members of the Club and its council could not have been any more en-

Number of Years ct

Robinson Fernald Rollins Hodgdon Bogle Tryon Nickerson DeWolf Conant Editors

re 3. Illustration oa the trend in decreasing tenure of Editors of Rhodora. When two Editor’s terms overlap each has been eauted off to the full yea

cl

elopoyuy

TOT ‘T°A]

1999] Sullivan—Centennial Reflections 13

couraging and supportive in the early years than they have been during my term thus far!

In addition to reflecting changes in the field of botany, Rhodora reflects the changes that have occurred in the membership of the New England Botanical Club. In its early years, the Club mem- bership consisted of an approximately equal mixture of amateurs and professionals; in a 1995 survey 75% of the respondents listed themselves as students or professionals having employment re- lated to botany. In addition, the early members of the Club were mostly ‘“‘resident’? members; that is, members who lived within 25 miles of Boston. By the time Fernald assumed the position of Editor-in-Chief, the journal boasted 33 Old World subscribers (Editorial 1929). Today the Club’s membership and subscribers range worldwide, and manuscripts are regularly submitted from outside of the U.S. One of the regular features of the journal, NEBC Meeting News, attempts to keep distant members in- formed of the content of our monthly seminars.

When he took on the job of Editor-in-Chief, Fernald began his term by publishing an overview of the accomplishments of the journal’s first 30 years (Editorial 1929). At that time, it was rea- sonable to relate such statistics as the number of new and total contributors per volume, and to name some of the more faithful contributors. Since then, many new, more specialized periodicals have begun publication, drawing papers away from the more gen- eral botanical journals such as Rhodora. Still, the journal has maintained publication of high quality papers on a variety of top- ics in botany. In addition, the elimination of page charges in 1996 has made publication in the journal more accessible to students and professionals with limited funds.

In his 1929 editorial announcement, Fernald took the oppor- tunity to restate quite eloquently the parameters within which manuscripts should fit in order to be appropriate for publication in Rhodora: “The pages of Rhodora are not reserved ... for members of the Club. They are freely open to all who care to use them, especially for the publication of tersely stated notes on

nge extensions or new or unrecorded facts regarding habits, morphology, habitats or other features of interest to students of

1 pl . or the natural history of plants. Systematic revisions and monographs of groups represented in the flora of northeastern North America will be welcomed for editorial consideration and well-written and descriptive (but not prolix) accounts of explo-

14 Rhodora [Vol. 101

rations, containing a good share of new or significant observa- tions, will be gladly considered. Mere lists without clear statement of the significance of the records are less desirable. Illustrations of new species and of newly recognized diagnostic characters are most desirable ... Photographs of landscapes, unless they are remarkably sharp vind of patent significance to the discussion, are undesirable for reproduction and, in general, Rhodora cannot commit itself to publish them. ... Manuscripts which show se- rious lack of consistency will necessarily be returned for correc- tion. In case of misquotations, erroneous citations and other in- accurate details in manuscripts the editors will naturally make corrections of obvious errors. They cannot, however, be expected to specially check such matters and it will be inferred that authors have themselves verified such essential details. Neither can the editorial board be held responsible . . . for all statements and con- clusions presented by different authors. In the case of controver- sial subjects, with the desire to present both sides of a question, papers may be accepted for publication, though not representing the views of the editors.’” While the possibilities may seem a bit limited by today’s standards, Fernald could not have anticipated the full range of submissions, especially the range of experimental techniques, available to researchers 70 years later.

The details of the discussions 100 years ago on the name of the new journal were not recorded in full. We know that Taxus was suggested in jest, and we know the other, serious consider- ations that were included in the vote. Apparently, some members felt that the name “Rhodora’”’ was “too sentimental,” perhaps because of the poem by Ralph Waldo Emerson, although that connection has not been mentioned elsewhere. Apparcotly, some members with more limited vision felt that the name ‘““Rhodora”’ would be appropriate for a club whose members had a primary interest in plants with the same range (Editorial 1929; Howard 1996; Pease 1951). Whatever the thoughts of those 26 voting members in 1898, Rhodora now serves readers worldwide and, while concentrating on the flora of North America, information on related taxa and comparable ecological phenomena from be- yond that limit are considered for publication.

ACKNOWLEDGMENTS. I am indebted to the members of the New England Botanical Club, especially the Council members and ed-

1999] Sullivan—Centennial Reflections 15

itorial staff, who have provided so much encouragement and sup- port during my three years as Editor.

LITERATURE CITED

Bessey, C. E. 1901. Baptisia tinctoria as a tumbleweed. Rhodora 3: 34—35.

BRAINERD, E. 1899. The saniculas of western Vermont. Rhodora 1: 7—9.

CanBy, W. M. 1900. Coreopsis involucrata on the Atlantic coast. Rhodora 2: 34

CHURCHILL, J. R. 1901. A botanical excursion to Mount Katahdin. Rhodora 3: 147-160.

COLLINS, - S. 1899a. Notes on algae.—I. Rhodora 1: 9-11.

———.. 1899b. A case of Boletus poisoning. Rhodora 1: 21-23.

1919. Chinese marine algae. Rhodora 21: 203-207.

CoLuns, J. E 1901. Notes on the bryophytes of Maine,—II. Katahdin mosses.

Rhodora 3: 181-184. Deane, W. 1899. A prolific fringed gentian. Rhodora 1: 11. EpIToRIAL. 1899. Editorial announcement. Rhodora 1: 1-2. . 190 3-284

———.. 1929. Editorial announcement. Rhodora 31: FERNALD, M. L. 1899. The ee. -plantains of New England. Rhodora ae Le

1. The vascular plants of Mount Katahdin. Rhodora 3: 166-177.

KENNEDY, G. G. AND J. E COo..ins. 1901. Bryophytes of Mount Katahdin. odora 3: 177-181.

MANNING, W. H. 1899. Matricaria discoidea in eastern Massachusetts. Rho- dora 1: 1

Pease, A. S. 1951 . The New England Botanical Club a half-century ago and later. Rhodora 53: 97-105.

RosInson, B. L. 1899. A new wild lettuce from eastern Massachusetts. Rho- ora 1: 12-13.

WessTER, H. 1899. Notes on some fleshy fungi found near Boston. Rhodora 1: 13-18.

WILLIAMS, E. E 1899. Myosotis collina in New England. Rhodora 1: 11-12. ———.. 1901. A comparison of the floras of Mt. Washington and Mt. Ka- tahdin. Rhodora 3: 160-165.

RHODORA, Vol. 101, No. 905, pp. 16-27, 1999

A REVIEW OF THE TAXONOMIC STATUS OF HACKELIA VENUSTA (BORAGINACEAE)

Ricuy J. HARROD AND LAurRI A. MALMQUIST

USDA Forest Service, Leavenworth Ranger District, 600 Sherbourne, Leavenworth, WA 98826

ROBERT L. CARR

Department of Biology, Eastern Washington University, Cheney, WA 99004

ABSTRACT. Morphological variables were sear using principal com- ponents and discriminant analyses to determine patterns of relationships among populations of Hackelia venusta, a narrow w endemic, and H. diffusa

with the population from the type locality at a low elevation clearly distinct from high elevation populations that have been assigned to this species. The high elevation populations represent an undescribed taxon. No affinities with either the low elevation H. venusta or the high elevation undescribed taxon were found to exist with populations of H. diffusa var. arida. Both H. venusta and the undescribed high elevation taxon are very narrow endemics and would benefit from well-developed conservation strategies and subsequent management.

Key Words: Hackelia, taxonomy, rare species

Hackelia venusta (Piper) St. John, showy stickseed, is a narrow endemic species of the Boraginaceae currently known only from Chelan County, Washington. As described by Gentry and Carr (1976), the species is a moderately stout perennial, 2—4 dm tall, often with numerous, erect to ascending stems from a rather slen- der taproot. It has large, white, showy flowers. The nutlets are 3— 4.5 mm long, with 8-14 intramarginal prickles. The marginal prickles are fused for up to % their length, forming a flange ca. 1 mm wide. It is found on steep, rocky slopes covered with gra- nitic scree.

The species was first described by Piper (1924) in the genus Sree and was later transferred to Hackelia by St. John (1929).

original description given by Piper was based on a 1920 eee made by J.C. Otis (895, Us) at a site about seven miles northwest of Leavenworth in Tumwater Canyon at an elevation

16

1999] Harrod et al.—Hackelia venusta 17

of 488 meters. Piper described H. venusta as having a white co- rolla about 2 cm broad. In 1947, a specimen (Long 14, ws) was collected about 16 km south, southwest of the Otis collection in the Alpine Lakes Wilderness, Chelan County, at an elevation of 2030 meters. Subsequently, researchers (Carr 1974; Gentry and Carr 1976; Hitchcock et al. 1959) included this alpine collection in their circumscription of the species and noted that flowers are white or sometimes washed with blue. Since that time, three ad- ditional alpine populations assumed to be H. venusta have been located, one from an area near the Otis collection (Harrod 238, Leavenworth Ranger District Herbarium), one from Asgaard Pass (plants have not been relocated since 1995, Harrod unpubl. data)

and the other from Cashmere Mountain, all above 2000 meters within the Alpine Lakes Wilderness area.

Some recent workers have suggested that the high elevation populations may be taxonomically distinct from the Tumwater Canyon Hackelia venusta (Gamon 1988; Loyal A. Mehrhoff, USFWS, Portland, OR, and Kathleen Robson, Robson Botanical Consultants, Vancouver, WA, pers. com.). The purpose of this study was to evaluate the relationship of these populations in order to achieve a better understanding of the taxonomic status of H. venusta. Because of the possibility of some allopatric in- trogression between H. venusta (sensu stricto) and populations of H. diffusa (Doug. ex Lehmann) Johnston var. arida (Piper) Carr in the lower end of Tumwater Canyon and several coulees north of Leavenworth (Carr 1974; Gentry and Carr 1976), a number of populations of the H. diffusa var. arida were included in the study.

MATERIALS AND METHODS

Study sites. Data were collected from ten populations in Washington shown on the map in Figure 1. Collection sites for Hackelia venusta (sensu lato) were located on the Wenatchee Na- tional Forest in Tumwater Canyon (TC), 9.6 km west of Leav- enworth, 488 meters; Crystal Creek (CC), 19.0 km southwest of Leavenworth, 2030 meters; and on Cashmere Mountain (CM), 16.0 km southwest of Leavenworth, 2073 meters. Collection sites for Hackelia diffusa var. arida were located on the Wenatchee National Forest in Tumwater Canyon (TW), 1.6 km west of Leav- enworth, 400 meters; Derby Canyon (DE), 11.3 km southeast of

18 Rhodora [Vol. 101

; Vv CRYSTAL CIRQUE A Low elevation H. venusta N V High elevation H. venusta A © H. diffusa var. arida

Figure 1. Locations of the populations of Hackelia examined in this study.

Leavenworth, 730 meters; Burch Mountain (BM), 4.8 km north- west of Wenatchee, 400 meters; Swakane Canyon (SC), 19.3 km northeast of Wenatchee, 1188 meters; and on the Ponderosa Es- tates Special Interest area (PE), 17.7 km north of Leavenworth, 670 meters. Two sites were located on Bureau of Land Manage- ment land; in Moses Coulee (MC), 24.0 km north of Quincy, 152 meters; and Douglas Creek (DC), 26.0 km north of Quincy, 140 meters.

Benson 02).

1999] Harrod et al.—Hackelia venusta 19

Morphological characters. Characters selected generally follow those used by Gentry and Carr (1976). Nineteen morpho- logical characters from three categories were scored for statistical analysis (vegetative, floral, and fruit) and an additional 11 qual- itative characters were recorded (Table 1). At each site, 25 plants were chosen randomly, numbered, and tagged. The Cashmere Mountain site, however, supports a small population and only 14 plants were selected. From each plant, one radial (basal) leaf and two cauline leaves, one from the lower one-third and one from the upper one-third of the stem, were chosen randomly for mea- surement. Three flowers and three fruits were chosen randomly and measured on each plant.

Statistical analyses. Both principal components and discrim- inant analyses were performed on the quantitative morphological data (SYSTAT 1997, SPSS Inc., Chicago, IL). Principal compo- nents analysis (PCA) was used to show natural groupings among each sampling unit or operational taxonomic unit (population). PCA is a method of partitioning a resemblance matrix into a set of perpendicular components (Ludwig and Reynolds 1988). Each component or axis has a corresponding eigenvalue which is the variance accounted for by that axis. The eigenvalues of the matrix are separated in descending order of magnitude so that each PCA component represents successively lesser amounts of variation (Ludwig and Reynolds 1988). The first component is the linear combination of variables accounting for more variance in the data than any other possible combination. The second component is the linear combination of the remaining variance after the first component is accounted for, the third component is the best linear combination after the first and second components have been ac- counted for, and so on. The data for the PCA involved the entire data set of a 238 X 19 character matrix (Table 1).

Discriminant analysis was used to establish the nonarbitrariness of group assignments. This analysis places each case within the group (population) with which it shares discriminating characters (Anderson and Taylor 1983). Unlike PCA, discriminant analysis is biased in that it positions cases within the ordination based on discriminating characters to achieve maximum separation of pre- viously defined groups. The case distributions were plotted by

two discriminant functions that separated the assigned groups to

Table 1. Morphological characters used in the taximetric analysis of Hackelia venusta and H. ae var. arida. All measure- ed

ments in mm unless otherwise no

Vegetative

Floral

Fruit

Plant height (dm)

Radial leaf length

Radial leaf width

Radial leaf petiole |

Radial leaf shape fet

Radial leaf surface (descriptive)

Lower cauline leaf length

Lower cauline leaf width

Lower cauline leaf shape (descriptive) Lower cauline leaf surface (descriptive) Upper cauline leaf length

Upper cauline leaf width

Upper cauline leaf shape (descriptive) Upper cauline leaf surface (descriptive)

Pedicel —

Calyx len

Calyx a (descriptive) Limb width

Corolla aa (descriptive) Anther len

Fornice Soak (descriptive) Fornice appendage height Fornice protuberance length

Nutlet shape (descriptive)

Nutlet surface (descriptive) Nutlet length

Number - intramarginal prickles Flange w

Distinct pickle ‘aug

Fraction connat

07

viopoyy

IOI ‘I°A]

1999]

Table 2. Means and standard deviation (in ee of characters used

in the present study. cept ns a tid. is

Measurements ar

Harrod et al.—Hackelia venusta

e given in mm,

in dm. ‘Abbreviations of the quantitative characters listed 1 in Table

Blue-flowered White-flowered or. cele Hackelia venusta Baca venusta Character! =2 = 1 Floral Ped 3.7 (1.42) 6.3 (1.88) 4.4 (1.84) Clx 3.0 (0.43) 3.8 (0.54) 2.4 (0.50) LimWid 4.2 (0.72) 7.4 (1.84) 4.3 (0.98) th 0.9 (0.13) 1.0 (0.15) 1.0 (0.52) For/Ap 1.0 (0.14) 1.3 (0.20) 0.6 (0.20) r/Pr 0.8 (0.19) 1.5 (0.31) 0.7 (0.32) Fruit NutL 5.6 (0.85) 6.4 (0.88) 6.2 (1.14) #InPr 10.2 (2.86) 11.4 (2.92) 10.1 (3.93) FIW 1.8 (0.34) 1.9 (0.38) 15 :(0:52) DPL 1.2 (0.25) 1.1 (0.76) 1.0 (0.53) FrCon 0.4 (0.11) 0.5 (0.08) 0.3 (0.12) Vegetative Height (dm) 1.4 (0.35) 2.7 (6.74) 5.1 (1.40) REL 56.9 (16.36) 48.9 (11.6) 98.6 (41.0) RL:W 14.4 (4.60) 11.3 (4.08) 8.8 (4.00) RL:Pet 21.9 (9.50) 32.2 (10.4) 63.3 (26.4) CLE 28.8 (6.64) 37:3 CEL.O) 83.2 (23.3) CLL:W 9.2 (2.85) 7.4 (2.00) 4.5 (1.62) CLUE 15.0 (5.67) 20.1 (6.50) a1 (13.3) CLU:W 6.5 (2.26) 6.6 (2.40) 3.7 (1.60)

the greatest ability. Again, the data for this analysis involved the same 238 X 19 character matrix used in the PCA.

Qualitative characters were not subjected to statistical analyses, but are used for further discussion and description.

RESULTS

The means and standard deviations for the quantitative char- acters are presented in Table 2 for each putative taxon. The Crys- tal Creek and Cashmere Mountain populations, which were blue- flowered, consistently had smaller floral measurements than the white-flowered Hackelia venusta of Tumwater Canyon. However, there were no consistent differences between these populations and the H. diffusa var. arida populations; there is considerable

22 Rhodora [Vol. 101

variability in floral size among populations of this latter, complex taxon. Fruit characteristics tended to be similar among all popu- lations. The Tumwater Canyon H. venusta were taller in stature than the Crystal Creek and Cashmere Mountain populations, but similar to all populations of H. diffusa var. arida that were ana- lyzed. Leaf characteristics were variable, with the Tumwater Can- yon H. venusta having the shortest radial leaves but intermediate in leaf length for the upper and lower cauline leaves. The Crystal Creek and Cashmere Mountain populations had the widest radial and lower cauline leaves.

Radial leaf and nutlet measurements were missing from a num- ber of cases at the conclusion of the study. These characters were dropped from both the principal components and discriminant analyses since the program would ignore those cases with missing data.

ponents that accounted for all the variance, the first three ac- counted for 68.0% (38.5%, 18.3%, and 11.2%, respectively). The

H. diffusa var. arida.

Finally, there was considerable overlap in the Hackelia diffusa var. arida cases with no distinct groups (Figure 2). However, there is some separation based on populations; the Swakane Canyon, for example, is separated from Ponderosa Estates and Derby Can- yon populations.

1999] Harrod et al.—Hackelia venusta 23

3 T T T Soi. sc DE sc DE 2 TW pc DE DE i om MSE D&c sis Tw sc DE pc E pc PE mc DE TC ae Mc BYE c pc DE DE @ sc BMsc p¢ tW TRE OME — | Sg aii mc_MC DE i 5 MC DEDE Be = sc me 5 Cc DE x Oo, vo ae : pETC aa BM PE ‘Say tw DC Mc, Miu rc TC ave Mc PE peer P Tc TC sc CMM PE Tc Cc TC cc ar PE PE . tye | p= cyec ce TC cc coc MS TC PECM GMC pe x< %. © ae cc RE TC ce ce orc PE 7c TG@c ome ay | | | | | -3 -2 -1 0 1 2 3 FACTOR(1)

Figure 2. Ordination of populations of Hackelia examined in this study based on scores of principal components | and 2. The first two components accounted for 56.8% of the total variance (38.5% and 18.3%, respectively).

diffusa var. arida: BM = Burch Mountain, TW = Tumwater Canyon; SC = Swakane Canyon, PE = Ponderosa Estates, MC = Moses Coulee, DE = Daly, Canyon, DC = Douglas Creek; H. venusta (white-flowered form): TC

= Tumwater Canyon; H. venusta (blue-flowered form): CM = Cashmere Mountain, CC = Crystal Creek.

Discriminant analysis. Table 3 gives the variables used and their relative usefulness in discrimination. The characters that contributed most, in order of importance, were height, fornice appendage height, fornice protuberance length, and limb width. Figure 3 shows the population centroids plotted on the basis of two (out of 9) of the most discriminating functions. Functions 1 and 2 accounted for 85.2% of the ability to distinguish amon groups (72.9% and 12.3%, respectively). The total peidinasbisity

24 Rhodora [Vol. 101

Table 3. Variables used in discrimination analysis #1 and their usefulness in discrimination among populations.

Function coefficients (+)

F (to Variable Function 1 Function 2 remove)

Pedicel length 0.019 0.158 5.42 Calyx length 0.271 0.189 2.78 Limb width 0.081 0.480 9.37 Anther len 0.016 0.013 +99 Fornice appendage height 0.604 0.304 34.23 Fornice protuberance length 0.259 0.654 18.45 Plant height 0.659 0.072 48.86 Upper cauline leaf length 0.278 0.264 6.52 Upper cauline leaf width 0.153 0.410 me A | Lower cauline leaf length 0.163 0.197 5.64 Lower cauline leaf width 0.386 0.228 Se

that a case from a certain population is correctly classified to that population was 81.0%. Predictability for the Cashmere Mountain and Crystal Creek populations was 82% and 76%, respectively, with individuals not showing affinity to each population grouping with the other. Only two individuals from the Cashmere Mountain population showed affinity to another population (Tumwater Can- yon, white-flowered Hackelia venusta). Ninety-two percent of cases were correctly classified in the Tumwater Canyon popula- tion. Predictability for the H. diffusa var. arida populations varied from 71% to 96% with deviant individuals grouping with other H. diffusa var. arida populations. The PCA showed some sepa- ration of the Swakane Canyon, Derby Canyon, and Ponderosa Estates populations, which is corroborated to some degree by the discriminate analysis (Figure 3). Predictability for the Swakane, Derby Canyon, and Ponderosa Estates populations was 96%, 83%, and 96%, respectively.

DISCUSSION

tinct from the high elevation collections which apparently rep- resent an undescribed taxon. We are in the process of completing

1999] Harrod et al.—Hackelia venusta 25

S zZ ” 4 C) nN «5 = -10 l [ae | -10 -5 0) 5 10 SCORE(1)

Figure 3. Ordination of populations of Hackelia examined in this study based on two most discriminating functions. Functions 1 and 2 spew for 85.2% of the ability to distinguish among populations (72.9% and 12.3% te H. diffusa var. arida: BM = Burch Mountain, TW = Tumwater

; SC = Swakane Canyon, PE = erener Estates, MC = Moses eee “DE = esnetd Canyon, DC = Douglas Creek; H. venusta (white- flowered form): TC = Tumwater Canyon; H. venusta (blue-flowered form): CM = Cashmere See, CC = Crystal Creek.

further studies on these and additional populations. The most ob- vious morphological distinction between the high elevation and Tumwater Canyon populations is flower color. The high elevation plants are always blue, while the Tumwater Canyon plants are largely white, sometimes with a faint blue tint. This study dem- onstrates that there are additional morphological distinctions, such as plant height, fornice appendage height, fornice protuberance length, and limb width. The high elevation and Tumwater Canyon populations also occupy markedly different environments, but

26 Rhodora [Vol. 101

both occupy similar substrate, scree derived from granodiorite and tonalite (Tabor et al. 1987). Additional factors considered include the absence of intermediate forms between the high and low elevation taxa and plants remain true to form and color when grown in a greenhouse (Harrod unpubl. data).

The results of our study do not suggest allopatric introgression between Hackelia venusta in Tumwater Canyon and H. diffusa var. arida as had been previously suggested by Gentry and Carr (1976). Some populations of H. diffusa var. arida do have larger flowers, but do not approach the size of the Tumwater Canyon H. venusta individuals. Other characters are also dissimilar. How- ever, allopatric introgression between H. venusta and H. diffusa var. arida as posed by Carr (1974) and Gentry and Carr (1976) can not be ruled out by our study since we found considerable variability in floral measurements among H. diffusa var. arida populations. Three populations (Swakane Canyon, Ponderosa Es- tates, and Derby Canyon) were separated from each other, but not from other populations of H. diffusa var. arida. The positions of Ponderosa Estates and Derby Canyon in the PCA and discrimi- nant ordinations were closer to H. venusta than any other popu- lations including Swakane Canyon (based largely on floral char- acteristics). However, it is unclear from our data whether or not gradation in floral characters within populations of H. diffusa var. arida are the result of allopatric introgression or simply site dif- ferences. More information is needed to discover this possible relationship.

Conservation concerns. The Tumwater Canyon Hackelia venusta consists of one small population with ca. 150 individuals located near a major state highway. The population in the early 1970s was estimated to occupy a few hundred acres (Carr 1974; Gentry and Carr 1976), but has dramatically decreased due to highway maintenance and habitat loss associated with fire exclu- sion and subsequent increase in woody vegetation, shading, and stabilization of scree slopes. This population could be lost due to random environmental events and, therefore, is severely threat- ened. In addition, the high elevation populations are also quite restricted and may be subject to loss from stochastic events. All three populations would benefit from well-developed conserva- tion strategies and subsequent management.

1999] Harrod et al.—Hackelia venusta 27

ACKNOWLEDGMENTS. We would like to thank John Gamon, Loyal Mehrhoff, and Kali Robson for their work showing the need for addressing this taxonomic problem. We appreciate the constructive comments Ted Thomas, John Gamon, Kali Robson, Loyal Mehrhoff, James Miller, and an anonymous reviewer pro- vided on early versions of this manuscript. Dottie Knecht, Mark Ellis, Cedar Drake, Ellen Kuhlmann, and Shelly Benson provided field assistance. We thank Pam Camp for help in locating popu- lations of Hackelia diffusa var. arida on BLM land. This project was cooperatively funded by the USFS, USFWS, and the Wash- ington Natural Heritage Program. Figure 1 was developed by Dan O’Connor, Wenatchee National Forest.

LITERATURE CITED

ANDERSON, A. V. R. AND R. J. TAYLOR. 1983. Patterns of morphological variation in a population of mixed species of Castilleja (Scrophulari- a2.

CARR, an es 1974. A taxonomic study in genus Hackelia in western North a. Ph.D. Dissertation, ei State University, Corvallis, OR.

GAMON, I. 1988. Habitat Management Guidelines for Hackelia venusta in the Wenatchee National Forest. Washington Natural Heritage Program, Olympia,

GENTRY, J. L. JR. AND R. L. Carr. 1976. A revision of the genus Hackelia (Boraginaceae) in North America, north of Mexico. Mem. New rk

Bot. Gard. 26: 121-227.

Hitcucock, C. L., A. CRONQUIST, M. OWNBEY, AND J. W. THOMPSON. 1959. Vascular Plants of the Pacific Northwest. Part 4: Ericaceae through Cam- anulaceae. University of Washington Press, Seattle, WA.

Lupwia, J. A. AND J. EF REYNOLDs. 1988. Statistical Ecology. John Wiley and Sons, New York.

Piper, C. V. 1924. New flowering plants of the Pacific Coast. Proc. Biol. Soc. Wash. 37: 91-96.

Sr. Joun, H. 1929. New and noteworthy northwestern plants. Res. Stud. State Coll. Wash. 1: 104-105.

Sane R. W,, V. A. FrizzeL, Jr., J. T. WHETTEN, R. B. Waitt, D. A. Swan-

NN, G. R. ByerLy, D. B. Bootu, M. J. HETHERINGTON, AND R. E. ZarT-

MAN. 1987. Geologic map of the Chelan 30-minute by 60-minute quad- rangle, Washington. Misc. Investigations Series, Map I-1661, U.S. Geol. Survey.

(Outs RHODORA, Vol. 101, No. 905, pp. 28-39, 1999

MYRIOPHYLLUM MATTOGROSSENSE (HALORAGACEAB), A RARE LOWLAND WATERMILFOIL NEW TO BOLIVIA

GARRETT E. CROW AND Nur P. RITTER Department of Plant Biology, University of New Hampshire, urham, NH 03824 (SOCOVOS ABSTRACT. Myriophyllum mattogrossense is reported as new to Bolivia.

This rare Watermilfoil of the Amazon Basin was previously known only from the original area of discovery in Brazil, one locality in the lowlands of Peru, and one in Ecuador. Notes on morphology, including a terrestrial growth form, and habitat are given, and a key is provided to differentiate the South American taxa of Myriophyllum.

Key Words: Myriophyllum, Haloragaceae, Watermilfoil, Bolivia

Since 1994 we have been conducting a broad biodiversity sur- vey of aquatic and wetland plants in Bolivia. While carrying out this fieldwork we encountered two small populations of Myrio- phyllum growing in streams ca. 20 km apart in the Amazon Basin region of Bolivia, known as the Chapare. The plants were found growing in swiftly flowing water of small rapids, rooted among rocks and gravel. These plants were conspicuously different from M. quitense Kunth (= M. elatinoides Gaud.), the common and widely distributed species in Bolivia. Although M. quitense is a common element of high elevation lakes, and is often so abundant that cattle are driven into the water to feed on it during the dry season (Dejoyx and Iltis 1991: Ritter and Crow 1998), we had not found any other populations below 2500 m and were sur- prised to encounter a Myriophyllum in the lowlands. Another spe- cies of Myriophyllum, M. aquaticum (Vell.) Verdc. (= M. brasi- liensis Camb.), is a widespread aquatic weed of tropical and warm

southern Brazil (Orchard, 1981). However in Bolivia, this species is known only from a newly discovered site in the Interandean

that our material was distinct from M. aquaticum.

We were ultimately able to determine the identity of the plant in the Chapare as Myriophyllum mattogrossense Hoehne, the first record known for Bolivia. Until recently, this rare species had been known only from two locations, one near Cuyaba, Mato

28

1999] Crow and Ritter—Myriophyllum mattogrossense 29

Grosso, Brazil, upon which E C. Hoehne (1915) based the de- scription for his new species, and one in the foothills on the eastern side of the Andes at Tocache Nuevo, Peru (Kahn et al. 1993; Orchard 1981). More recently, M. mattogossense was col- lected from a third location, near Coca, Ecuador (Orchard and Kasselmann 1992). Orchard (1981) noted that the species might well be found eventually in a much wider area of the lower foot- hills on the eastern side of the Andes of Peru, Brazil, and perhaps even Bolivia, and attributed the lack of known sites to the sub- merged habit and inconspicuous flowers.

Moreover, it is our experience that aquatic plants, in general, are greatly undercollected in the Neotropics. Many aquatic plants which are rather common are poorly represented in herbaria. Ad- ditional populations of Myriophyllum mattogrossense surely exist, but are not likely to be encountered unless the fieldwork is spe- cifically focused on aquatic plants. This was certainly the case when Christel Kasselmann, a specialist of aquatic plants for aquarium culture, collected the first record for Ecuador (Orchard and Kasselmann 1992). We stumbled onto the first Bolivian pop- ulation while searching for members of the Podostemaceae, an aquatic family restricted to rapids and swift flowing waters in areas with a seasonal fluctuation of water levels. Thus, M. mat- togrossense is now known from its type locality in Mato Grosso, Brazil, and in the Amazon Basin near the base of the Andes in Ecuador, Peru, and Bolivia (Figure 1).

The Chapare region, where the Bolivian populations were en- countered, borders the eastern slope of the Andes and is notable for having the highest amount of rainfall in Bolivia, with parts of the region receiving more than 5000 mm of precipitation per year (Ribera et al. 1994). The larger rivers and tributaries of the area experience a high level of disturbance during the rainy sea- son. River courses in the Chapare are extremely transitory, with riverbeds receiving large depositions of gravel and sand, and with new channels frequently being formed while former stretches are transformed into curiches (oxbows). Streams and other tributaries can also experience significant disturbance as well. Generally speaking, the streams in the area are characterized by a lack of rooted vegetation and haptophytes (Crow and Ritter, pers. obs.). In the case of Myriophyllum mattogrossense, it appears that a combination of fairly specific habitat requirements—clear, fast- moving water and a substrate composed of gravel and cobbles—

30 Rhodora [Vol. 101

Figure 1. Documented distribution of Myriophyllum mattogrossense.

coupled with the transitory nature of aquatic habitats due to se- vere disturbances, serves to limit the number of populations of this species.

Previously, this species was believed to be strictly submersed. Orchard (1981) stated that the species is unusual in that its flow- ers and fruits are unusually small, and that the plant was reported to grow completely submersed, resulting in underwater opening and pollination of the flowers. While we observed the submersed

1999] Crow and Ritter—Myriophyllum mattogrossense 31

plants to be fertile, as did Hoehne (1915), we also observed the existence of a terrestrial growth form for Myriophyllum matto- grossense, likewise in fertile condition. The terrestrial growth orm was initiated as the water level dropped and marginal plants became stranded (Figure 2). The submersed leaves dried up and new upright branches sprouted from the prostrate stem. When seen in this condition, the species had an almost moss-like, or Hippurus-like appearance (Figure 2). Although this was observed in both of the Bolivian populations, there was no mention of a terrestrial growth form on the labels of the Peruvian specimens examined. However, Kasselmann (Orchard and Kasselmann 1992) observed emergent plants growing on mud along the riv- erbank, which fit the description of the terrestrial growth form we observed.

In the Bolivian material the leaves of submersed plants have segments that, while filiform, are very thin, distinctly flattened, with a conspicuous midvein, and which are wider than typical for Myriophyllum quitense. The Peruvian material examined exhib- ited the same morphology. We were able to examine only one herbarium specimen of the Ecuadorian material, and while the submersed leaves were flattened, the segments were much more filiform than those of either the Bolivian or Peruvian material. However, they did closely resemble those depicted in the illus- tration accompanying Hoehne’s (1915) original description, now serving as the lectotype (Orchard 1981). Previously, we had noted that the markedly capillary leaf segments in the Brazilian popu- lation were altogether distinct from those of the other populations. We were able to reconcile this variation by attributing it to habitat differences (lacus temporarius in Brazil). Arber (1920) noted that water plants respond to certain physical stimuli and that in My- riophyllum, in particular, one can observe marked differences in the morphology of the same species growing in different current regimes. In still water, plants may have leaf segments that are delicate and nearly hair-like, while the stresses of current on the leaves of plants growing in strongly flowing water require that leaves tend toward increased mass and thickness (Arber 1920; Gerber and Les 1994).

In contrast to the submersed plants, the leaves of the terrestrial form have divisions that, while still somewhat flat, are pectinate (with fewer divisions), thicker, and distinctly succulent (Figure

32 Rhodora [Vol. 101

Figure 2. Habit of terrestrial growth form of Myriophyllum mattogros- sense at edge of stream.

Figure 3. Close-up view of terrestrial growth form of Myriophyllum mat- togrossense showing somewhat flattened, thicker, succulent, pectinate leaves.

1999] Crow and Ritter—Myriophyllum mattogrossense 33

3). The flowers and fruits are axillary on both terrestrial and sub- mersed plants.

DESCRIPTION OF BOLIVIAN MATERIAL

Plants perennial, herbaceous aquatics, with submerged and ter- restrial growth forms (Figure 4). Stems and leaves with small sessile glands; glands moderately dense on young growth, becom- ing sparse on older growth. Submersed growth form: stems flex- uous, ascending in quiet water, somewhat horizontal in flowing water; leaves verticillate, in whorls of 3—4, pinnately divided, ca. (18—-)20—22(—25) mm long, with 7-8 pairs of lateral segments (mostly alternate), segments flattened, 0.4—-0.5 mm wide, each with a distinct midvein; hydathodes filiform, tiny, present at base of petioles and each leaf segment on young growth. Terrestrial growth form: stems of submersed plants rooting on stream mar- gins or gravel bars, submersed leaves withering away; upright stems arising from axillary buds, not flexuous, sturdy, erect; leaves verticillate, pectinate, mostly 10-11 mm long, becoming shorter toward stem tip (4—5 mm long), mostly with 3 pairs of lateral segments (alternate), segments somewhat flattened, thick- ish, slightly succulent, each with a distinct midvein (especially on herbarium material) each segment with an apical secretory gland; hydathodes filiform, tiny, present at base of petioles and each leaf segment on young growth. Flowers (both growth forms) axillary, 1—4 per whorl, bisexual, appearing sessile (pedicel short, 0.25—0.4 mm long), subtended by a pair of bracteoles (apparently early caducous), frequently with filiform hydathodes on each. Perianth 4-merous, opposite the ovary lobes, alternate with sta- mens. Stamens 4, subsessile, anthers ovoid, slightly apiculate at tip, stamens developing before stigmas, not long persisting. Ova- ry inferior, 4-lobed, stigmas 4, conical; tiny hydathodes present at summit of ovary. Fruits globose, 4-lobed, 7-9 mm long, 7-9 mm wide; mericarps with a few weak tubercles on outer surface.

Flowering in this species did not appear to be seasonal. Based on all specimens examined, flowering material has been observed on specimens collected in February, March, April, May, June, and Nov

Orchard and Kasselmann (1992) noted a number of features evident in the Ecuadorian populations which had not previously been observed in Myriophyllum mattogrossense, thus expanding

34 Rhodora [Vol. 101

Figure 4. Myriophyllum mattogrossense drawn from submersed growth form specimens. (A) Section of stem showing axillary flowers, glandular emergences (appearing as dots) scattered on leaves and stems, and hydathodes present at leaf bases. (B) Section of young shoot. (C) Young flower with only stamens evident. (D) Mature flower with stigmas alternating with stamens, and with subtending bracteole present.

1999] Crow and Ritter—Myriophyllum mattogrossense ee)

the description for the species. These features were, in particular, the presence of filiform ‘‘hydathodes,” the trichomas collectores of Hoehne (1915), at the bases of the petioles and at the base of each leaf segment on young growth; the presence of numerous, scattered, globular sessile glands on the surface of the young stems and leaves; and flowers with a complete absence of a peri- anth. The Bolivian material is consistent with all of these features with exception to that of the perianth. In the Bolivian specimens the flowers do possess a single, 4-merous perianth whorl of small triangular appendages, arranged alternate the styles and opposite the 4 stamens, the stamens developing first and not persisting. A further character we noted was the presence of a pair of bracteoles subtending the flowers (Figure 4), which apparently are caducous, as they were noted only with earlier stages of flowers. Myrio- phyllum mattogrossense had previously been described as lacking bracteoles (Orchard 1981; Orchard and Kasselmann 1992); a lack of bracteoles is unusual in the family (Orchard and Kasselmann 1992)

The presence of sessile glands is unusual in the genus (Orchard and Kasselmann 1992), thus the feature can serve as a good di- agnostic character for Myriophyllum mattogrossense. Since glands had not been noted on the Peruvian material (Orchard 1981), we re-examined the Peruvian herbarium specimens; glob- ular sessile glands are, indeed, present.

Recently, some puzzling reports of Myriophyllum mattogros- sense in the Gran Pantanal of Mato Grosso, Brazil, have appeared in the literature. Prado et al. (1994) noted that M. mattogrossense forms “luxuriant beds” during the high water stages in the Pan- tanal. This species was said to “‘bloom intensively” during this time, and then to die off. The authors stated that M. mattogros- sense is “‘easily recognized in the field by its deep red, densely clustered leaves,” and further noted that the species possesses emergent flowers which descend below the water’s surface fol- lowing fertilization (Prado et al. 1994, p. 581). Clearly, red veg- etation and emergent inflorescences are characters not known to be associated with M. mattogrossense. Unfortunately, the identity of their plants cannot be confirmed as no voucher specimens had been cited.

Heckman (1997) reported Myriophyllum mattogrossense as fill- ing the niche of submersed plants in the tropical wet-and-dry climatic zone in South America. In a subsequent book on the

36 Rhodora [Vol. 101

Brazilian Pantanal, he described ‘‘luxuriant submerged beds” of M. mattogrossense that form in the northern Pantanal during the high water stage, and included a color photograph of the pre- sumed M. mattogrossense (Heckman 1998). Having examined this photograph, we have concluded that the species in question is clearly not M. mattogrossense.

Although Pott and Pott (1997) included Myriophyllum matto- grossense in their comprehensive checklist of aquatic plants of the Brazilian Pantanal, they noted that they have never observed this species in the Pantanal, and were aware of its presence only through the original type collections of Hoehne (Vali Pott, pers. com., 1998). Furthermore, Guarim Neto’s (1992) checklist of an- giosperms of the Pantanal includes no species of Myriophyllum. In like manner, during our extensive expedition in 1998 in the Bolivian portion of the Pantanal we encountered neither M. mat- togrossense nor any other species of Myriophyllum.

SPECIMENS EXAMINED

fast-moving water, 5 May 1996, Ritter 3147 (LPB, Mo, NHA).

Brazil. Mato Grosso, near Cuyabd. Original specimens of EF C. Hoehne apparently lost (Orchard 1981). Lectorype: Tabula n. 127 (“Ns. 4.578 e 4.635. Hab. lacus temporarius ad Coxip6 da Ponte, propre Cuyabé”), Comm. Linh. Telegr. Mato Grosso Amaz., Annexo 5, Bot. 6. 1915.

Ecuador. Rio Coca, 8 Feb 1990, Kasselmann 133 (p). According to Or- chard and Kasselmann (1992) the site locality is: Rio Yanaucu (lower Rio Coca drainage) about 20 km north of the town of Coca (Pto. Francisco de Orellana) at the crossing of the road from Coca to Lago Agrio.

- Department of San Martin: Province of Mariscal C4ceres, al oeste de vivero del Instituto Agropecuario de Tocache, Tocache Nuevo, elev. 400 m, 10 Nov 1969, Schunke V. 3598 (GH, US).

Peru. Department of San Martin, Province of Mariscal Caceres, Fundo Jeroglifico, del Sr. Luis Ludefia (Quebrada de Ishichimi), Tocache Nuevo, elev. 400 m, 10 Apr 1975, Schunke V. 8281 (MO).

Peru. Department of San Martin: Province of Mariscal Caceres, Quebra- da Ishichimi, cerca al Fundo del Sr. Luis Ludefia, sumergida en las riachuelos, elev. ca. 400 m, 3 Nov 1980, Schunke V. 12393 (Mo).

1999] Crow and Ritter—Myriophyllum mattogrossense 37

In order to facilitate differentiation of the species of Myrio- phyllum in South America the following key is provided, includ- ing information from our expanded understanding of M. matto- grossense and M. quitense (Ritter and Crow 1998). Additionally, it is noteworthy that although M. spicatum L. has been listed for Peru and M. verticillatum L. has been listed for Chile, specimens bearing those names were based on misidentified specimens (Or- chard, 1981); these two species are not known to occur in South

KEY TO THE SOUTH AMERICAN SPECIES OF MYRIOPHYLLUM

1. Submersed leaves in whorls of (2—)3—4(—5); segments of sub- mersed leaves very slender, distinctly flattened, ca. 0.5 mm wide, with conspicuous midvein, or filiform; emergent leaves absent on submersed form in reproductive phase (terrestrial growth form, with ascending aerial branches sprouting from prostrate stems, may be expected to occur stranded along water margin; leaves pectinate, mostly 10— 11 mm long, segments mostly 3 per side); leaves and stems bearing scattered, small, globular, sessile glands (es- pecially young material); flowers solitary, borne axillary along submersed portion of stem, (also axillary along erect stems on terrestrial form); flowers bisexual; stamens 4;

mericarps with a few weak tubercles on outer surface; rare, lowlands, Ecuador, Peru, Bolivia, and Brazil ...... fie GaN A ieee E.G eck s Ok Sie oe tok ea A M. mattogrossense

1. Submersed leaves in whorls of (3—)4—6; segments of sub- mersed leaves filiform to only somewhat flattened, mostly up to ca. 0.25 mm wide, midvein not conspicuous; emer- gent leaves present on submersed form (terrestrial form rare in M. quitense, ascending aerial branches sprouting from prostrate stems, leaves pectinate, mostly 7-10 mm long, segments 5—6(—8) per side); leaves and stems lacking glands; flowers in spicate inflorescences, borne in the axils of the emergent leaves only; flowers unisexual, plant mon- oecious or dioecious (bisexual flowers on terrestrial form in M. quitense); stamens 8; mericarps smooth on outer a ls ie seals nahh o ba Lbns $9 os Kh ae we de bes Z

2. Submersed leaves in whorls of (3—)4(—5), ovate in outline, 1—2 cm long, with 7-9 pairs of pinnae, segments nearly

38 Rhodora [Vol. 101

filiform, somewhat flattened; emergent leaves blue- green, tinted red or purple, in whorls of (3—)4, ovate to oblong, more or less entire, at least in upper parts, toothed to pinnatisect in lower parts; plants monoecious (flowers bisexual in terrestrial form); Andes from Ven- ezuela to Tierra del Fuego, e. Argentina, s. Uruguay, and Falkland Islands, disjunct tae eee in Mexico, nw. North America, PE.I., Canada ..... M. quitense 2. Submersed leaves in whorls of rence 6, oblanceolate in outline, (1.7—)3.5—4 cm long, with 12-15 pairs of pin- nae (lower leaves decaying rapidly); distinctly filiform, terete; emergent leaves glaucous, in whorls of (4—)5— 6, narrowly oblanceolate, pectinate, with (9—)12—18 pairs of pinnae; plants dioecious (female plants only in adventive populations); s. Peru, s. Bolivia, and s. Brazil south to c. Chile, n. Argentina, and pe eh intro- duced weed northward in Mesoamerica and e. I oS eC Oe Pe M. cant

ACKNOWLEDGMENTS. We are thankful to Carol Morley for pre- paring the botanical illustration. Robynn Shannon kindly provid- ed us with a photocopy of Hoehne’s original description. We are grateful to Christel Kasselmann for providing us with a living specimen of Myriopohyllum mattogrossense from Ecuador. Drs. Thomas D. Lee and A. Linn Bogle reviewed a draft of the man- uscript and provided helpful comments. The curators of the fol- lowing herbaria are acknowledged for loans of herbarium speci- mens: B, GH, MO, and Us. This research was supported in part by a grant from the Vice President for Research and Public Service, University of New Hampshire. This paper is Scientific Contri bution Number 1978 from the New Hampshire Agricultural oo periment Station.

LITERATURE CITED

ARBER, A. 1920. Water Plants: A Study of Aquatic Angiosperms. Reprint edition with preface by W. T. Stearn. 1972. J. Cramer. Lehre, Germany.

Desoyx, C., AND A. ILtIs. (eds). 1991. El Lago Titicaca, Sintesis del Cono- cimiento Limnolygico Actual. ORSTOM, Institut Francais de Recherche Scientifique por le Development en Cooperation. La Paz, Bolivia.

GerseR, D. T. and D. H. Les. 1994. Comparison of leaf morphology among

1999] Crow and Ritter—Myriophyllum mattogrossense 39

submersed species of pe es niculio ee eee subs different hab- i r. J. Bot. 973-979

HEckMAN, C. A. 1997. Ecoclimatological survey of the wetland biota i Me the tropical wet-and-dry climatic zone. Global Ecol. Biogeogr. Lett. 6: 97— 114.

———. 1998. The Pantanal of Pocone—Biota and Ecology in the Northern Section of the World’s Largest Pristine Wetland. Kluwer Academic Pub- lishers. Dordrecht, The Netherlands.

Hoeune, FE C. 1915. Commissao de Linhas Telegraphicas, eg ona a Mato Grosso ao Amazonas. Annexo No. 5. Historia Natural Ri il

Kaun, F, L. BLANCA, AND K. R. YOUNG. (compiladores). 1993. Las Plantas Vasculares en las Aguas Continentales del Pert. Inst. Francés de Estudios Andios. ear Peri

ORCHARD, A 1981. A revision of South American Myriophyllum (Halor- agac ceae), er its repercussions on some Australian and North American — Brunonia 41: 27-65.

. KASSELMANN. 1992. Notes on Myriophyllum mattogrossense alragacse, cae J. Bot. 12: 81-84.

Pott, V. J. AND A. Pott. 1997. Checklist do macr6fitas acudticas do Pantanal, pad Acta Bot. Brasil. 11: 215-227.

Prapo, A. L. po, C. W. HECKMAN, AND FE B. Martins. 1994. The seasonal succession ‘of biotic communities in wetlands of the tropical wet-and-

dry climatic zone: II. The aquatic macrophyte vegetation in the Pantanal of Mato Grosso, Brazil. Int. Rev. Gesamten Hydrobiol. 79: 569-589.

RIBERA, M. O., M. LIEBERMAN, S. BECK, AND M. Mora ces. 1994. Mapa de la Vegetacién y Areas Protegidas de Bolivia. Instituto de Ecologia. La Paz, Bolivia.

Ritter, N. P. AND G. E. Crow. 1998. Myriophyllum quitense H.B.K. (Hal- oragaceae) in Bolivia: A terrestrial growth-form with bisexual flowers.

saa atic Bot. 60: 389-3

. In pr | Primera Se pee de Myriophyllum aqua- ticum ties en Bolivia. Ecol. Bolivi

RHODORA, Vol. 101, No. 905, pp. 40-45, 1999

DISTRIBUTION AND DENSITY OF SUBMERGED AQUATIC VEGETATION BEDS IN A CONNECTICUT HARBOR

Topp A. RANDALL

Gulf Coast Research Laboratory, P.O. Box 7000, Ocean Springs, MS 39566

JOHN K. CARLSON

University of Mississippi, Department of Biology, University, MS 38655

MATTHEW E. MrRoczKA

Cedar Island Marina Research Laboratory, PO. Box 181, Clinton, CT 06417

ABSTRACT. Submerged aquatic vegetation (SAV), Zostera marina and

Key Words: SAV, distribution, Long Island Sound, Clinton Harbor

Submerged aquatic vegetation (SAV) of the North American Atlantic coastal waters supports highly diverse animal assem- blages (Heck et al. 1989; Rozas and Odum 1987). Seagrass beds function as refugia, energy sources, and habitat for the animals inhabiting the beds (Heck et al. 1989; Sogard and Able 1991). A decline in seagrass bed production would have profound effects upon these animal assemblages and would decrease detrital ex- port, greatly reducing energy sources for other fauna not directly inhabiting SAV beds (Thayer et al. 1984).

The reduction of SAV in the coastal United States has been well documented (Kemp et al. 1983: Orth and Moore 1983; Rob- blee et al. 1991; Thayer et al. 1994). Natural causes of SAV decline such as disease, storm events, salinity fluctuations, and

40

1999] Randall et al—Submerged Aquatic Vegetation 41

hypoxic events coupled with anthropogenically induced eutro- phication currently threaten the production of many SAV com- munities (Durako and Kuss 1994; Koch and Beer 1996; Monta- gue and Ley 1993; Olesen and Sand-Jensen 1994; Zieman et al. 1994). Therefore, documenting the distribution of SAV is impor- tant in developing baseline data which can be used to monitor abundance patterns and ecological health over extended periods of time.

The distribution of SAV (Zostera marina L. and Ruppia mar- itima L.) along the Connecticut coast of Long Island Sound has been previously documented by Koch and Beer (1996). However, detailed maps of the extent of seagrass beds within individual bays and harbors are not available. Historically, SAV has been reported to occur in Long Island Sound as far west as New York State but is now limited to the easternmost third of Long Island Sound (Koch and Beer 1996). Clinton Harbor is considered the westernmost distribution point of seagrass in Long Island Sound (Koch and Beer 1996) and was selected as a study site to monitor changing SAV distribution patterns. The purpose of this study was to provide baseline information on the density and the dis- tribution of SAV in inner Clinton Harbor (Clinton, CT).

Clinton Harbor occupies 162 ha and is located on the Con- necticut shore of Long Island Sound. The harbor is a drowned river valley inundated by seawater and receives freshwater input from the Hammock, Indian, and Hammonasset Rivers. Inner Clin- ton Harbor is the mouth of the Hammonasset River and is formed by the presence of the Cedar Island spit (Figure 1). The tides within the harbor are semi-diurnal and display a mean tidal range of 1.5 meters. Clinton Harbor is a rural harbor with 60% of its bordering land edge being utilized as wetlands and beaches, 25% as marinas, and 15% as residential housing (Mroczka 1991).

MATERIALS AND METHODS

SAV densities were measured in inner Clinton Harbor from August 15, 1990 through October 31, 1990. Inner Clinton Harbor was divided into 33 transects set at 30 m intervals along the shoreline. The transect positions were established with survey equipment, marked with stakes, and subsequently plotted on a hydrographic survey map.

SAV densities were obtained using SCUBA. Divers moved

INNER CLINTON HARBOR

METERS

100

200

Figure 1. Map of inner Clinton Harbor indicating the location — density of submerged aquatic vegetation. Density categories are expressed as the mean number of short shoots per m? (Low =

O0—20; Medium

= 21-40; High = 41).

eIOpoyy

IOI TOA)

1999] _—_ Randall et al—-Submerged Aquatic Vegetation 43

along a calibrated 200 m line making observations at 15 m in- tervals. At each interval, SAV short shoot densities were counted in two randomly placed 0.5 m? grids. Zostera marina and Ruppia maritima short shoots were not differentiated in the counts. Counts were adjusted to m? values and the mean of the two counts was then calculated and used as a datum point for mapping. A total of 662 sampling points, each sampled once, was used in this study. SAV distribution was plotted on a survey map (Figure 1) using the density categories established from percent shoot cov- erage of a m’ grid. The densities established were categorized as low (O—20 short shoots per m*), medium (21—40 short shoots per m7’), and high (=41 short shoots per m7). Area covered (ha) was then calculated for each density category with a Scalex Planwheel planimeter.

RESULTS

SAV was documented in an estimated 23 ha of inner Clinton Harbor. Although Zostera marina and Ruppia maritima shoots were not differentiated in the counts, the beds were dominated by Z. marina. Nine areas of low density SAV beds were identified (Figure 1). The low density beds were distributed throughout the study area, but were located primarily in areas north of the nav- igational channel. Low density beds occupied approximately 11 ha. Nine areas of high density SAV beds, occupying appro mately 6.5 ha, were located laterally along the navigational chan- nel (Figure 1). Ten areas of medium density beds were also iden- tified laterally along the navigational channel (Figure 1). Medium density beds accounted for approximately 5.5 ha of all SAV beds. The mudfiat along the northern shore of the inner harbor was unpopulated by SAV, as was the navigational channel.

DISCUSSION

Seagrass beds play integral roles in coastal ecology and have been suggested to be among the most productive aquatic ecosys- tems known (Day et al. 1989). Extensive declines in SAV along the coastal United States in the past few decades have made it necessary to document and monitor the extent of existing seagrass beds (Koch and Beer 1996; Olesen and Sand-Jensen 1994). Base- line data of SAV distribution will provide researchers and re-

44 Rhodora [Vol. 101

source managers with the necessary information to begin to rec- ognize and interpret the effects of both natural and anthropogenic impacts on the seagrasses, which will be critical for the future management of this resource.

The results of our 1990 survey have shown that inner Clinton

Harbor was dominated by low density seagrass beds (Figure 1), which were located mainly in the areas north of the navigational channel. This portion of the harbor is dominated by poorly sorted mud/silt or (Mroczka 1991) and is typically exposed during peri rmal low tide. High turbidity significantly reduces light aeaiabitny and the production of the beds throughout this area. High and medium density SAV beds were identified in large aggregations laterally along the navigational channel (Figure 1). High flow rates of water through the navigational channel keep the surrounding sediments composed of well sorted sand grains (Mroczka 1991). High water flow also reduces the turbidity in the areas surrounding the channel making the water considerably clearer than that of the mudflats to the north. Water clarity in these areas permits the critical light level to extend to the bottom, allowing for more productive seagrass beds (Day et al. 1989; Duarte 1991).

The lack of SAV within the navigation channel is due to the depth of the channel and periodic oyster dredging. The 3 m depth of the navigational channel does not allow sunlight to penetrate to a level that is conducive to SAV growth (Duarte 1991). Peri- odic oyster dredging in the channel also precludes the establish- ment of SAV beds.

Our study provides baseline data on the general distribution of SAV in inner Clinton Harbor. Future SAV monitoring can be conducted and subsequently compared to the results of this sur- vey in order to determine changes in the distribution patterns of the seagrass beds.

ACKNOWLEDGMENTS. We thank Jeffrey Shapiro and Dr. Peter Pellegrino for their support of this project. Thanks are also ex- tended to Stephen Capella and Kimberly Damon for their assis- tance in SAV counts. We are grateful to J. D. Caldwell, Cynthia Moncreiff, and Robin McCall for help with the manuscript prep- aration and review.

1999] Randall et al—Submerged Aquatic Vegetation 45

LITERATURE CITED

Day, J. W., C. A. S. HALL, W. M. Kemp, AND A. YANEZ-ARANCIBIA. 1989. Estuarine Bestey. Wiley and Sons, New York.

DuartE, C. M. mie Seagrass ng ter limits. Aquatic Bot. 40: 363-377.

Durako, M. J. AND K. M. Kuss. 1994. Effects of Labyrinthula infection on the nanan tien ace of Thalassia testudinum. Bull. Mar. Sci. 54: 727-732

Heck, K. L., K. W. ABLE, M. P. FAHAY, AND C. T. ROMAN. 1989, Fishes and decapods of Cape Cod eelgrass meadows: Species composition, seasonal abundance patterns and comparison with unvegetated substrates. Estu- aries 12: 59-65.

Kemp, W. M., R. R. TwiLtey, J. C. _ STEVENSON, W. R. BOYNTON, AND J. C. MEANS. Wig The decline of submerged vascular plants in upper Ches- apeake B Yee e of results concerning possible causes. J. Mar.

85.

Kocu, E. W. AND . "BEER. 1996. Tides, light and distribution of Zostera marina in Long Island Sound, USA. Aquatic Bot. 53: 97—107. MontTaaug, C. L. AND J. A. Ley. 1993. A possible effect of salinity fluctuation on abundance of benthic vegetation and associated fauna in northeastern Florida Bay. Estuaries 16: 703-717.

Mroczka, M. E. 1991. An investigation of the finfish utilization of a co ne marina basin with special emphasis on the winter flounder (Pseudop ronectes americanus Walbaum). M.S. thesis, Southern Connecticut pine

, New Haven, CT.

OLESEN, B. AND K. SAND-JENSEN. 1994. Patch = see of eelgrass Zostera marina. Mar. Ecol. Progr. Ser. 106: 147-15

OrtH, R. J. AND K. A. Moore. 1983. eae Bay: An unprecedented decline in eee aquatic vegetation. Science 222: 51-52.

ROBBLEE, ee B., T. R. BarBER, P. R. Cartson, M. J. DurAKo, J. W. Four- se , L. K. MUEHLSTEIN, D. Porter, L. A. YARBRO, R. T. ZIEMAN,

Diss Lc. ZIEMAN. 1991. Mass mortality of the tropical seagrass Thal- es testudinum in Florida Bay (USA). Mar. Ecol. Progr. Ser. 71: 297-

209.

Rozas, L. P. AND W. E. Opum. 1987. The role of submerged aquatic vege- tation in influencing the abundance of nekton on contiguous tidal fresh- water marshes. J. Exp. Mar. Biol. Ecol. 114: 289-300.

SocarD, S. M. AND K. W. ABLE. 1991. A comparison of eelgrass, sea lettuce macroalgae, and marsh creeks as habitats for epibenthic fishes and deca-

s. Es 19.

Tuayer, G. W., W. J. KENworTHY, AND M. S. Fonseca. 1984. The ecology of eelgrass meadows of the Atlantic coast: A community profile. USFWS FWS/OBS-84/02.

MurPHEY, AND M. W. LACrorx. 1994. Responses of plant com- nuunities in western Florida Bay to the die-off of seagrasses. Bull. Mar. Sci. 54: 718-726.

ZIEMAN, J. C., R. DAvIs, J. W. FOURQUREAN, AND M. B. RosBLEE. 1994. The

role of climate in the Florida Bay seagrass dieoff. Bull. Mar. Sci. 54: 1088

RHODORA, Vol. 101, No. 905, pp. 46-86, 1999

THE DISTRIBUTION OF THE BRYOPHYTES AND VASCULAR PLANTS WITHIN LITTLE DOLLAR LAKE PEATLAND, MACKINAC COUNTY, MICHIGAN

C. Eric HELLQUIST! AND GARRETT E. CROW Department of Plant Biology, University of New Hampshire, Durham, NH 03824 ‘ Current Address: 391 West Road, Adams, MA 01220

ABSTRACT. Little Dollar Lake peatland, a Sphagnum-dominated poor fen peatland complex, has a flora consisting of 36 bryophyte and 93 vascular plant species. Random, quantitative sampling of 279 one-meter-square quad- rats along 13 transects on six Lars and mats was analyzed by two-way indi- cator species analysis (TWINSPAN). Based on interpretation of the TWIN- SPAN analysis, three vegetation cover types and six constituent vegetation phases were delineated. The three cover types were designated as the Cala- magrostis canadensis cover type (lagg habitats), the Chamaedaphne calycu- lata cover type (peatland mat habitats), and the Chamaedaphne calyculata— Triadenum fraseri cover type (transitional habitats with evidence of terres- trialization). A fourth association, the Potamogeton confervoides—Utricularia Scop cover type, was recognized based on qualitative field observa-

pera evidence suggests that terrestrialization and small-scale paludification are occuring in some areas of the peatland.

Key Words: peatland, bog, fen, Sphagnum, bryophytes, vascular plants, ter- restrialization, TWINSPAN, Michigan

Peatlands are wetlands especially characteristic of northern re- gions across North America and Eurasia that form in cool, tem- perate or maritime climates where evapotranspiration is low. Peat deposits dominate the boreal environments of Canada (170 x 10° ha), the former Soviet Union (150 X 10° ha), and the United States (36.8 X 10° ha; Gorham 1990). Near the southern edge of the glacial boundary in North America, peatlands tend to be more sporadic and are usually confined to smaller areas (typically gla- cial scour basins or kettleholes), as opposed to the extensive mires that blanket the landscape of subarctic latitudes (Crum 1988). Along the glacial boundary, peatland basins are frequently frost pockets that serve as refugia for northern plants at the southern edge of their ranges (e.g. Andreas and Bryan 1990; Crum 1988; Damman and French 1987; Pielou 1979).

The peat of boreal regions consists of partially decomposed

46

1999] Hellquist and Crow—Little Dollar Lake Peatland 47

remains of bryophytes (especially Sphagnum) and vascular plants (primarily sedges and ericaceous shrubs), as well as minute amounts of inorganic matter (Crum 1988; Damman and French 1987; Moore and Bellamy 1974). Although the majority of wet- lands produce peat to some extent, northern peatlands store vast quantities of peat because rates of organic deposition greatly ex- ceed rates of decomposition (Crum 1988).

Peatlands are typically associated with acidic water chemistry. The main causes of peatland acidity are the result of a variety of biogeochemical processes and positive feedbacks (Crum 1988; Gorham 1957; van Breeman 1995) involving the presence of or- ganic acids within the anoxic subsurface peat (Hemond 1980; Kilham 1982), the high cation exchange capabilities of Sphagnum taxa (Andrus 1986; Clymo 1963, 1964; Crum 1988; Kilham 1982; Spearing 1972; Vitt et al. 1975), the uptake of ions by vegetation (Kilham 1982), the overall hydrological characteristics of individual peatland systems (Kilham 1982; Vitt et al. 1975), the oxidation of reduced sulfur Seaside (Clymo 1964; Gor- ham 1961), and, occasionally, acid precipitation (Clymo 1964; Crum 1988).

Conditions within peatlands are generally considered less than optimal for vascular plant growth and establishment (Clymo and Hayward 1982; Moore and Bellamy 1974; van Breeman 1995; Vitt et al. 1995). However, extensive bryophyte communities thrive in peatland environments of boreal landscapes (Vitt 1990). Due to their intimate contact with the aqueous chemical environ- ment of the peatland, bryophytes are strong indicators of micro- habitat conditions (Vitt and Slack 1984). Peatland vegetation dis- pean itself based largely on gradients of minerotrophy (e.g.

kalinity) and pH, as well as light and hydric conditions of mi- crohabitats (e.g. Anderson and Davis 1997; Gignac and Vitt 1994; Jeglum 1971; Vitt and Chee 1990; Vitt and Slack 1975; Wheeler et al. 1983).

Peatlands exhibit strong floristic similarities across much of North America, resulting in similar species composition patterns in regions such as the Northeast and upper Midwest of North America (Crum 1988; Gore 1983; Wheeler et al. 1983). From site to site, the diversity of both bryophyte and vascular peatland flo- ras is highly influenced by the microhabitat heterogeneity of the peatland. The greater the diversity of microhabitats available, the

48 Rhodora [Vol. 101

greater the species richness of the peatland (Anderson and Davis 1997; Vitt et al. 1995)

An extensive and often confounding literature exists regarding the classification of peatland systems. Researchers classify peat- lands utilizing a variety of terms and definitions that assess peat- land attributes ranging from edaphic and physical characteristics to vegetation community composition (Gore 1983). Based on wa- ter chemistry, peatlands can be classified generally into two polar groups, Oombrotrophic peatlands (bogs) and minerotrophic peat- lands (fens). A third group, oligotrophic peatlands, have inter- mediate water chemistry and floristic characteristics.

Ombrotrophic peatlands receive all mineral subsidies from at- mospheric deposition or from nutrients released during the slow decay of organic matter within the peatland basin. Thus, these peatlands are extremely nutrient-limited (Crum 1988; Damman and French 1987; Gore 1983). Ombrotrophic peatlands may be cut off from drainage or flushing by their geomorphological ori- entation, such as in a kettlehole bog. These peatlands may also be isolated from the water table by their own prolific deposition of peat that can elevate the peatland above the influence of groundwater, such as in a maritime raised bog (Crum 1988; Dam- man and French 1987). Due to the lack of active water circula- tion, hydrogen ions accumulate in ombrotrophic peatlands often resulting in pH values of 4.01—4.25 in the narrow definition of Gorham and Janssens (1992a), or less than 5.0 in the broader definition of Vitt et al. (1995). Bryophyte communities in om- brotrophic and oligotrophic peatlands are extensive and are dom- inated by Sphagnum species (Crum 1988; Gorham and Janssens 1992a; Janssens and Glaser 1986; Schwintzer 1981; Vitt 1990, Vitt and Slack 1975, 1984; Vitt et al. 1995)

Minerotrophic peatlands (fens) are relatively nutrient-rich com- pared to their ombrotrophic counterparts. Fens receive mineral subsidies from rheotrophic (flowing) ground or surface sources, as well as from precipitation. The nutritive input of a fen is unique from site to site based on the physical orientation and hydrology of the peatland. Thus, the mineral concentrations of fens are tremendously variable (Damman and French 1987; Gig- nac and Vitt 1994; Gore 1983). Edaphic inputs enrich fen en with higher concentrations of calcium, iron, magnesium, alumi num, and phosphorus ions and higher relative pH levels ons) ombrotrophic peatlands (Gignac and Vitt 1994; Gorham 1990;

1999] Hellquist and Crow—Little Dollar Lake Peatland 49

Jeglum 1971; Schwintzer 1978; van Breeman 1995). Fens are often subclassified as poor fens (pH 4—6) to extreme rich fens (pH 6—7.5) based on the relative minerotrophy, pH values, and plant species composition (pH values for the upper Midwest from Crum

The relative minerotrophy and higher pH levels of fens are reflected in their floras which become more species-rich along a gradient from low to high nutrient availability, i.e., oligotrophic to minerotrophic conditions (Crum 1988; Damman and French 1987; Glaser 1987; Gorham 1990; Schwintzer 1981). Fens typi- cally have an extensive sedge cover that is often dominated by Carex lasiocarpa Ehrh., a less prominent bryophyte cover, and greater overall species richness compared to ombrotrophic peat- lands (Crum 1988; Glaser 1987; Schwintzer 1978; Vitt et al. 1995)

Bryophytes in fens tend to be dominated by members of the Amblystegiaceae, including Amblystegium spp., Calliergon spp., and Scorpidium scorpioides (Hedw.) Limpr. (Gorham and Jan sens 1992a; Janssens and Glaser 1986; Vitt 1990; Vitt et ] 1995). Sphagnum taxa have reduced frequencies in fen ecosys- tems, but are nevertheless present. Species such as Sphagnum teres (Schimp.) Angstr. ex C. Hartm. and Sphagnum subsecundum Nees ex Sturm are typically prominent in alkaline fen conditions, particularly on the edge of the mat along open water (Crum 1988).

Most basin peatlands containing ponds or lakes exhibit the classic zonation of vegetation communities where a floating com- munity of sedges, Sphagnum, and/or ericaceous shrubs encroach over the open water of a lake (e.g., Crow 1969; Dunlop 1987; Fahey 1993; Fahey and Crow 1995; Schwintzer and Williams 1974). The more or less concentric vegetation patterns of peat- lands that radiate outward from lake margins led early ecologists to propose the “‘hydrosere model” of peatland succession. hydrosere model states that lake margin communities fill in or terrestrialize the open water of a pond or lake, and initiate a successional sequence that passes through ericaceous shrub as- sociations, coniferous forest associations, and finally culminates in an upland forest climax community on what was formerly the peatland (e.g., Dansereau and Segadas-Vianna 1952; Gates 1942; Transeau 1903).

Recently, this traditional hydrosere explanation that links spa-

50 Rhodora [Vol. 101

tial zonation with successional processes has been called into question by peatland ecologists (Klinger 1996; van Breeman 1995). Although the process of terrestrialization in basin peat- lands has been well documented, Klinger (1996) has reservations regarding two major tenets of the hydrosere model. First, it is unlikely that a mesic forest “‘climax”’ is the ultimate successional destiny of a peatland basin. Second, terrestrialization of peatlands is not necessarily a unidirectional process; instead it is quite often a dynamic progression of vegetation advance and recession along a water body (e.g. Schwintzer and Williams 1974).

Citing a lack of data from stratigraphic, dendrochronological, and vegetational analyses to support strict hydrosere peatland suc- cession, “the bog climax model” of peatland succession has been proposed (Klinger 1996). This model states that peatland condi- tions are their own climax in middle to high latitudes. In this model, peatland mats extend outward from lake margins via ter- restrialization, while Sphagnum mosses along the outer edge of the peatland expand beyond the limits of the basin and into the upland via paludification. Therefore, in the absence of large scale disturbances and alterations of peatland hydrology, peatland sys- tems remain ecologically intact for thousands of years (Klinger 1996). Early in peatland succession allogenic factors are impor- tant, but they later become secondary to autogenic factors that are initiated and maintained by the flora itself (Klinger 1996).

This paper outlines the vegetation associations found within Little Dollar Lake peatland based on field observations and TWINSPAN analysis. Comparisons of vegetation patterns at Lit- tle Dollar Lake peatland to other North American peatlands in the upper Midwest, Canada, and the Northeast are emphasized. The role of terrestrialization and paludification in peatland suc- cession is also discussed with reference to vegetation patterns observed at Little Dollar Lake peatland.

SITE DESCRIPTION

The Little Dollar Lake peatland basin is 14 hectares (34 acres) in area. The peatand is located in west-central Mackinac County on the eastern Upper Peninsula of Michigan (T44N, R8W, NE1/ 4 Sec. 28, Hudson Township). Little Dollar Lake has acidic water chemistry (mean pH = 4.5, n = 40) and is situated in a shallow glacial scour basin. Based on pollen stratigraphy and sedimenta-

1999] Hellquist and Crow—Little Dollar Lake Peatland 51

Outlet

Northwestern Mat

ey, Little Dollar Lake

Western Mat

Southern Mat

Southwestern Mat Extension

40m

0

Figure 1. Map of Little Dollar Lake peatland illustrating transect loca- tions and selected features of local geography around and within the peatland

Upland islands on southwestern mat; 3. Upland peninsula; 4. Muck pools in the southern mat extension; 5. Stream channel; 6. Location of the transect on the eastern mat parallel to the upland.

tion, the peatland around Little Dollar Lake began to form a proximately 3500 years before present as a result of prolonged increases in the regional water tables of eastern upper Michigan (Futyma 1982). Basin morphology, post-glacial history, soils, sur- rounding upland vegetation, and climate are described by Hell- quist (1996) and Hellquist and Crow (1997).

Little Dollar Lake is surrounded by seven peatland mats of varying extent. These mats were named based on their compass- point position around Little Dollar Lake (Figure 1). Each mat is colonized by a variety of plant associations. In general, the peat-

52 Rhodora [Vol. 101

land is characterized by a nearly continuous layer of Sphagnum spp. covered by expanses of open ericaceous scrub. A narrow, floating mat fringes much of the lakeshore and a graminoid lagg borders the upland of several mats.

MATERIALS AND METHODS

Bryophyte and vascular vegetation sampling. Vegetation analysis and mapping were conducted using transect sampling procedures. Thirteen transects, intersecting as many plant com- munities as possible, were surveyed and divided into ten-meter intervals. Using a table of pseudo-random numbers, two one- meter-square (1 m X 1 m) quadrats were chosen for sampling within each ten-meter interval. In some spatially narrow habitats, extra quadrats were added to help insure adequate sampling. From July 12, 1995 through August 15, 1995, 279 one-meter-square quadrats were sampled visually for frequency and absolute per- cent cover of bryophyte and vascular species. Cover was defined

s ““... an estimate of the area of coverage of the foliage of the species in a vertical projection on to the ground” (Shimwell 1971, p. 110). Species composition and absolute percent cover were estimated separately for the bryophyte and vascular strata in each quadrat.

SPAN vegetation analysis. Bryophyte and vascular percent cover data from the 279 sampled quadrats were analyzed using two-way indicator species analysis (TWINSPAN; Hill 1979). TWINSPAN is a polythetic divisive procedure that uses reciprocal averaging to ordinate quadrats, ultimately creating an ordered site-by-species two-way table (Hill 1979). Seven pseu- dospecies cut levels (Hill 1979: van Tongeraan 1987) were es- tablished as follows: 1 (0O-1% cover), 2 (1-2% cover), 3 (2-5% cover), 4 (5S—10% cover), 5 (10-25% cover), 6 (25-50% cover), and 7 (50—100% cover). All cut levels were weighted equally.

For studies involving the classification of ecological commu- nities, there are several advantages to employing TWINSPAN analysis. These include its use of raw data, its hierarchical clas- sification of both plots and species, and the ability to rewrite the arranged data matrix in a dendrogram format that enhances the clarity of plot relationships (Gauch 1982). TWINSPAN analysis has been utilized in many investigations of peatland vegetation

1999] Hellquist and Crow—Little Dollar Lake Peatland 33

Table 1. Species richness of the four vegetation cover types (CT) and six cover phases (PHS). The number of quadrats in each cover type or phase is noted in parentheses

Species

TWINSPAN Community Delineation Richness Potamogeton confervoides—Utricularia geminiscapa CT 7 Calamagrostis canadensis CT (27) 54 Iris versicolor—Lycopus uniflorus PHS (13) 44 Sphagnum cuspidatum—Dulichium arundinaceum PHS (14) 30 Chamaedaphne calyculata CT (220 56 Sphagnum recurvum—Carex oligosperma PHS (186) 53 Sphagnum magellanicum—Sarracenia purpurea PHS (34) 27 Chamaedaphne calyculata—Triadenum fraseri CT (32) ST Sphagnum majus PHS (22) 26 Sphagnum papillosum PHS (10) a

(Anderson and Davis 1997; Dunlop 1987; Fahey and Crow 1995; Miller 1996; Slack et al. 1980; Vitt and Chee 1990; Vitt et al. 1990)

RESULTS

TWINSPAN classification. Field observations and data weighed in concert with TWINSPAN analysis resulted in the de- lineation of four distinct vegetation cover types (including the aquatic vegetation of the lake) and six constituent vegetation phases within the Little Dollar Lake basin (Table 1; Figures 2 and 3). The aquatic cover type was recognized based on field pee Although these cover types were delineated pri- ough TWINSPAN analysis, only plant communities that could fe eed clearly in the field have been recognized. The complete TWINSPAN two-way table and raw data are available in Hellquist (1996). Those species that were prominent both in the field and in the SPAN analysis were chosen as appropriate species to name the cover types (CT) and phases (PHS). Due to overlap of indi- cator species between clusters, or to the relatively inconspicuous nature of the TWINSPAN indicator species in the soli tend of a cover type or phase, the TWINSPAN indicator species w not always adopted as appropriate “representative species’ core the names of cover types and phases. Indicator species for the cover types and phases are noted in Tables 2-10. Although sim-

279 252 27 CALAMAGROSTIS CANADENSIS 220 32 13 14 CHAMAEDAPHNE CALYCULATA CHAMAEDAPHNE CALYCULATA- Iris versicolor- Sphagnum cuspidatum- TRIADENUM FRASERI Lycopus uniflorus Dulichium arundinaceum 107 ii3 22 10 Sphagnum majus Sphagnum papillosum 34 73 Sphagnum magellanicum- Sphagnum recu Sarracenia purpurea Carex pr case

Figure 2. Dendrogram of the vegetation cover types and phases as delineated by TWINSPAN and field observation. Designations in all capital lettering are cover types, designations with capital and lowercase lettering are phases. Numbers refer t of quadrats in each cover type or phase. The arrow indicates that the clusters of 113 and 73 plots were combined to form the Sphagnum recurvum—Carex oligosperma phase

vs

eviopoyuy

TOT TOA]

1999] Hellquist and Crow—Little Dollar Lake Peatland 55 CALAMAGROSTIS CANADENSIS COVER TYPE Iris versicolor-Lycopus uniflorus Phase

q Sphagnum cuspidatum-Dulichium arundinaceum Phase

CHAMAEDAPHNE CALYCULATA COVER TYPE Sphagnum recurvum-Carex oligosperma Phase Sphagnum magellanicum-Sarracenia purpurea Phase CHAMAEDAPHNE CALYCULATA-TRIADENUM FRASERI COVER TYPE

wa Sphagnum majus Phase

(SS . ?

Ss

Sy

Figure 3. Vegetation map of Little Dollar Lake peatland illustrating the extent of the vegetation cover types and constituent vegetation hases. le Sphagnum papillosum phase is not mapped due to its narrow spatial distri- bution. This phase is present in a narrow (~ 1.0 m) band along the lake margin in all areas where the Sphagnum magellanicum—Sarracenia purpurea phase

is present. Scale is approximate.

56 Rhodora [Vol. 101

ilarities to other peatland complexes in the upper Midwest do exist, the names of these communities are not intended to have regional applicability due to the inherent variability of hydrology and topography that directly influences the composition of indi- vidual peatland floras.

INSPAN classification split the initial set of 279 quadrats into two groups, a group of 252 quadrats and a second group of 27 quadrats (Figure 2). The 27 quadrats were designated as the Calamagrostis canadensis CT. At the second level the C. cana- densis CT was divided further into phases of 13 and 14 quadrats respectively. The cluster of 13 quadrats was named the Iris ver- sicolor—Lycopus uniflorus PHS, and the cluster of 14 quadrats was named the Sphagnum cuspidatum—Dulichium arundinaceum PHS (Figure 2).

The remaining 252 quadrats were split into two clusters, one cluster of 220 quadrats and one cluster of 32 quadrats. The cluster of 220 quadrats was named the Chamaedaphne calyculata CT or “ericaceous scrub,” and the cluster of 32 quadrats was named the C. calyculata-Triadenum fraseri CT (Figure 2). The two phases of the C. calyculata CT consisted of clusters of 186 quad- rats and 34 quadrats.

TWINSPAN separated the 220 quadrats of the Chamaedaphne calyculata CT into clusters of 107 and 113 quadrats (Figure 2). The cluster of 113 quadrats at the third level had an essentially

bined into a cluster of 186 quadrats (Figure 2). These 186 quad- rats were representative of the Sphagnum recurvum—Carex oli- gosperma PHS (hummock-hollow complex) of the C. calyculata By

The remaining 34 quadrats, isolated from TWINSPAN’s cluster of 107 quadrats at the fourth level, had a unique suite of species that corresponded to the floating-mat community along the north- ern and southern lake margin. This cluster was named the Sphag- num magellanicum—Sarracenia purpurea PHS of the Chamae- daphne calyculata CT (Figure 2).

Lastly, at the second cut level, the cluster of 32 quadrats that composed the Chamaedaphne calyculata-Triadenum Sraseri CT was divided into two constituent phases (Figure 2). The Sphag-

1999] Hellquist and Crow—Little Dollar Lake Peatland 57

num majus PHS consisted of 22 quadrats and was associated with the stream channel and the narrow eastern and western mats (Fig- ures 1 and 3). The remaining ten quadrats, the Sphagnum papil- losum PHS, formed an approximately two to three meter zone immediately bordering the open water of the majority of the peat- land.

Species richness and frequency within sampling quad- rats. The Calamagrostis canadensis CT and the Chamaeda- phne calyculata CT were the most species rich of the cover types (Table 1). The C. canadensis CT was restricted to more miner- otrophic lagg areas. Despite being sampled by only 27 quadrats, this cover type contained 54 species. The C. calyculata CT was slightly more diverse, but was sampled extensively by 220 quad- rats. The most diverse phase was the Sphagnum recurvum—Carex oligosperma PHS of the C. calyculata CT (Table 1). The next most diverse phase was the Jris versicolor—Lycopus uniflorus ae of the C. canadensis CT (Table 1). The 7. versicolor—L. unifloru PHS consisted of only 13 quadrats but contained 44 different species. The most species-poor cover type or phase was the Po- tamogeton confervoides—Utricularia geminiscapa CT. With only seven species, this cover type was restricted to the open water of Little Dollar Lake.

The nutrient-poor nature of the peatland was emphasized by the dominance of the Sphagnaceae (peat moss family) and the Ericaceae (heath family) on a visual and a quantitative level with- in the 279 sampling quadrats. Of the ten most frequent species in the 279 quadrats, eight belonged to either the Sphagnaceae or the Ericaceae (Figure 4). Of these ten species, Chamaedaphne calyculata (L.) Moench was the most abundant species, occurring in 258 quadrats (92%). Vaccinium oxycoccos L. was second, oc- curring in 218 quadrats (78%). The most abundant sedge was Carex oligosperma Michx. which was present in 106 quadrats (38%).

The tenth most frequent taxon in the 279 quadrats was Acer rubrum L. This species was present only as seedlings or small saplings in the quadrats sampled. Seedlings were noted in 23% of the quadrats sampled. In every quadrat in which A. rubrum appeared, it had a cover value of less than 2%. Of the ten most abundant species, the only bryophytes were members of the Sphagnaceae (Figure 4

58 Rhodora [Vol. 101

Percent frequency from 279 quadrats wa i) |

» ° rs se ew aw Ny P 40 a0 < x y) “ i se Fol a ss Ao ox &

we ro 4

6 * g Vv e< x e e RS 9

Figure 4. Percent frequency of the ten most abundant species in the 279 quadrats sampled. The number above each bar equals the total number of quadrats in which each species occurred.

DISCUSSION

The vegetation of Little Dollar Lake peatland. The bryo- phyte flora of Little Dollar Lake peatland consisted of 36 species including eleven species of Sphagnum. The vascular flora of the peatland was comprised of 93 species dominated by sedge species (Cyperaceae) and heath species (Ericaceae). The complete bryo- phyte and vascular flora of Little Dollar Lake peatland, including comments on abundance and habitats, is presented in Hellquist and Crow (1997).

1999] Hellquist and Crow—Little Dollar Lake Peatland 59

The following discussion summarizes the vegetation patterns at Little Dollar Lake peatland, and emphasizes the presence of various plant species as a means to infer the nutrient status of a peatland or microhabitats within a peatland. The four cover types and the six constituent phases are discussed in a roughly centrip- etal manner starting with the lagg communities on the outside of the peatland basin and proceeding inward toward Little Dollar Lake.

1. Calamagrostis canadensis Cover Type

The outermost portion of a basin peatland is known as the lagg or moat. The lagg is an ecotone (sensu Risser 1995) between the consolidated peat of the open mat and the mineral soils of the upland. The lagg is characterized by shallower, better aerated peat that is enriched by nutrients from the adjacent upland (Crum 1988; Damman and French 1987; Gore 1983). A moat of open water separating the upland from the peatland proper is often found within the lagg area. The lagg is typically one of the most botanically diverse communities within a peatland. The lagg usu- ally contains minerotrophic wetland species that are not exclu- sively associated with peatland floras except in the context of lagg habitats (Crum 1988). Thus, the species composition of the lagg reflects a more minerotrophic or marsh-like physiognomy com- pared to the majority of the peatland basin.

The Calamagrostis canadensis CT formed an encircling, fen- like community adjacent to the upland on all peatland mats of the basin except the narrow eastern and western mats (Figure 3). In northern Michigan, a C. canadensis (Michx.) P. Beauv. asso- ciation typically follows fire and sometimes develops in the mar- ginal areas of peatlands that are damp, but not extremely wet (Gates 1942). In the summers of 1994 and 1995, this cover type was merely damp with no standing water. In 1996, however, most of this cover type was saturated with water or had standing water 0.25 to 0.50 m in depth, especially in areas of the southwestern mat and southern mat extension.

Sphagnum species were not as prominent in the oeyiene canadensis CT, with only S. recurvum P. Beauv. and S. datum Hoffm. occurring with some frequency (Table 2). ied num recurvum was apparent in the open, outer areas of lagg closer to the ericaceous scrub, while S. cuspidatum formed lush pockets in the wettest, muckiest areas of the lagg. In 1996, S. cuspidatum

60 Rhodora [Vol. 101

Table 2. Mean percent cover and percent frequency of dominant and as- sociated species in the Calamagrostis canadensis cover type consisting of 27 quadrats out of 279. Species having less than 10% frequency are not included

e.

in the tabl ‘AN indicator species. Lycopus uniflorus was also des- ignated as an indicator species, but does not appear in the table. Mean % Cover % Frequency BRYOPHYTES Sphagnum recurvum 55 a1L9 Sphagnum cuspidatum* 27 30.0 Warnstorfia fluitans 5 14.8 Calliergon cordifolium 27 fT Calliergon stramineum 22 ia Callicladium haldanianum 4 li VASCULAR PLANTS Calamagrostis canadensis* 30 88.9 Carex lasiocarpa 38 74.1 Potentilla palustris 5 55.6 Triadenum fraseri 2 dit Chamaedaphne calyculata 18 37.0 Acer rubrum <l 37.0 Lysimachia thyrsiflora 2 30.0 Iris versicolor* 14 26.0 Glyceria canadensis 26 13.5 Lysimachia terrestris 2 18.5 Galium trifidum A 15.0 Dulichium arundinaceum 20 14.8 Impatiens capensis 8 14.8 ospe 10 11.) Scutellaria galericulata 5 Ll Equisetum fluviatile 2 EEA

flourished in the standing water of southern lagg areas with del- icate individuals growing to lengths as long as 0.25 m. While there was a lack of Sphagnum diversity in this cover type, mem- bers of the Amblystegiaceae contributed to the diverse bryophyte flora of this cover type. Members of this minerotrophic-indicative family, Calliergon cordifolium (Hedw.) Kindb. and C. strami- neum (Brid.) Kindb., often were observed thriving in wet, decom- posing leaf litter.

The two most frequent and dominant vascular species were Calamagrostis canadensis and Carex lasiocarpa (Table 2). These two species grew intermingled and gave the lagg community its narrow-leaved graminoid texture that was readily distinguished

1999] Hellquist and Crow—Little Dollar Lake Peatland 61

from the neighboring ericaceous scrub community. At the height of the growing season, the prominence of C. canadensis gave the lagg a distinct marsh-like appearance. Other widely scattered, but locally abundant species of this community type included Galium tinctorium L., Iris versicolor L., Lysimachia thyrsiflora L., L. ter- restris (L.) B.S.P., Lycopus uniflorus Michx., and Potentilla pal- ustris (L.) Scop.

The species within the graminoid lagg at Little Dollar Lake peatland were similar to the herbaceous component of the lagg at Mud Pond Bog (Moultonborough, NH). Species found in both of these peatlands included Calamagrostis canadensis, Lycopus uniflorus, Scirpus cyperinus (L.) Kunth, and Scutellaria galeri- culata L. (C. E. Hellquist, ms. in prep.). The lagg communities at Mud Pond Bog were also characterized by an extensive tall shrub lagg association dominated by J//ex verticillata (L.) A. Gray, Nemopanthus mucronatus (L.) Trel., Lyonia ligustrina (L.) DC., Vaccinium corymbosum L., and Viburnum nudum L. (C. E. Hell- quist, ms. in prep.). At Little Dollar Lake, typical lagg shrubs such as /. verticillata and N. mucronatus grew infrequently within the peatland basin itself, although both taxa grew abundantly in rich upland soils that immediately fringed the lagg. The presence of I. verticillata, N. mucronatus, and V. nudum vat. cassinoides (L.) T. & G. at Little Dollar Lake is reminiscent of the /. verti- cillata-N. mucronatus community type described for northern Michigan kettlehole peatlands (Vitt and Slack 1975). This cover — also resembled the I. verticillata-Acer—Carex canescens co t found nearest the upland in Mud Pond Bog (Hillsborough. NH; Dunlop 1987). Species common to these cov- er types include J. verticillata, Acer rubrum, Lysimachia terres- tris, Lycopus uniflorus, and C. canadensis (Dunlop 1987).

1A. Iris versicolor-Lycopus uniflorus Phase

The substrate of the /ris versicolor-Lycopus uniflorus PHS (Figure 3) was composed of very shallow peat (<1.0 m) that often was covered by decaying leaf duff from upland trees. The presence of shade-tolerant Sphagnum squarrosum Crome empha- sized the nutrient-rich nature of the lagg habitat. Sphagnum squarrosum typically grows in minerotrophic, alkaline Thuja oc- cidentalis L. swamps and swampy woodlands (Andrus 1980; Crum 1983, 1988). The most abundant non-Sphagnum bryophyte was Calliergon cordifolium, a species that thrived in damp leaf

62 Rhodora [Vol. 101

Table 3. Mean percent cover and percent frequency of dominant and as- sociated species in the Iris versicolor-Lycopus uniflorus phase of the Cala- magrostis canadensis cover type (13 quadrats). Species having less than 10% frequency are not included in the table. *TWINSPAN indicator species.

Mean % Cover % Frequency

BRYOPHYTES Sphagnum recurvum 60 84.6 Warnstorfia fluitans a 15.4 Calliergon cordifolium* 27 23-1 Callicladium haldanianum 4+ 235.1 Drepanocladus uncinatus 6 15.4 Dicranum flagellare fi 15.4 Pleurozium schreberi 8 15.4

VASCULAR PLANTS Calamagrostis canadensis 44 100.0 Carex lasiocarpa 31 G9:2 Potentilla palustris o 61.5 Lycopus uniflorus ey 61.5

— ge i 38 g. g 2 is) 3 > _ A ms — Ann Wd Wo oo oO

Triadenum fraseri 3

Lysimachia terrestris 2 38.5 Impatiens capensis 8 30.8 Galium trifidum 3 30.8 Chamaedaphne calyculata 6 231 Scutellaria galericulata 5 23.1 Carex canescens 8 15.4 Carex oligos, 13 15.4 Viola macloskeyi 6 15.4

well as the mosses Climacium dendroides (Hedw.) Web. & Mohr.

peatlands may have a fringing Jris association less than a meter wide (Gates 1942). At Little Dollar Lake, I. versicolor was a

1999] Hellquist and Crow—Little Dollar Lake Peatland 63

prominent component of the C. canadensis CT with a distribution several meters in width, especially in the lagg areas of the south- ern-oriented peatland mats.

Within the Jris versicolor—Lycopus uniflorus PHS there was one locality that had an entirely unique species composition com- pared to the rest of the peatland. This area was situated within the peNe i of the southwestern mat (Figure 1) in an area that ap-

ently was influenced by runoff from a seasonal, dirt truck trail (Heliquist 1996). In this area, Typha latifolia L. was well estab- lished. Futyma (1982) noted the presence of Typha in the basin in the early 1980s, but made no reference to the location of the colony. Other species essentially limited to this distinct lagg hab- itat were Carex stipata Muhl., Epilobium ciliatum Raf., Polygo- num cilinode Michx., Rubus canadensis L., Scutellaria laterifiora L., and Rumex obtusifolius L.

1B. Sphagnum cuspidatum—Dulichium arundinaceum Phase

The Sphagnum cuspidatum—Dulichium arundinaceum PHS oc- curred in two distinct areas in the southeastern and western lagg areas of the southern mat extension (Figure 3). This association was apparent in the wettest areas of lagg that contained exposed, mucky peat. In 1996, this area was covered by standing water roughly 0.25 to 0.50 m deep. The most frequent vascular species of this phase included Calamagrostis canadensis, Carex ad carpa, Chamaedaphne calyculata, and Potentilla palustris. Thirtee species with greater than 10% frequency were present Sona Glyceria canadensis (Michx.) Trin., Carex utriculata F Boott, Equisetum fluviatile L., and Salix pedicellaris Pursh (Table 4).

aie dominant peat mosses of this phase were Sphagnum re- curvum and S. cuspidatum (Table 4). Other important bryophyte species included members of the Amblystegiaceae such as Warns- torfia fluitans, W. exannulata (Schimp. in B. S.G.) Loeske [Dre- panocladus exannulatus (Schimp. in B.S.G.) Warnst.], Calliergon cordifolium, and C. stramineum. Despite its infrequency in the quantitative sampling of this phase, W. exannulata formed an almost homogeneous bryophyte cover on exposed, mucky peat in the southeastern lagg of the southern mat extension. In 1996, this same population of W. exannulata was still vigorous despite being submerged in standing water. Warnstorfia exannulata is considered a dominant species of poor fens in Alberta along with Sphagnum angustifolium (C. Jens. ex Russ.) C. Jens., S. majus

64 Rhodora [Vol. 101

Table 4. Mean percent cover and percent frequency of dominant and as- sociated species in the Sphagnum cuspidatum—Dulichium arundinaceum phase of the Calamagrostis canadensis cover type (14 quadrats). Species having less than 10% frequency are not inched in age table. *TWINSPAN indicator species.

Mean % Cover % Frequency BRYOPHYTES Sphagnum recurvum 38 21.4 Sphagnum cuspidatum* 27 S71 Warnstorfia fluitans 8 14.3 Sphagnum maju | 14.3 Calliergon stramineum 30 14.3 Calliergon cordifolium 27 jf By | VASCULAR PLANTS Calamagrostis canadensis* 14 78.6 Carex lasiocarpa* 43 78.6 Chamaedaphne calyculata 24 50.0 Potentilla palustris S, 50.0 Lysimachia thyrsiflora* 1 me | Triadenum fraseri <1 35:1 Glyceria canade 32 28.6 Dulichium arundinaceum* 20 28.6 Acer rubrum ca | 21.4 Carex utriculata 8 14.3 Vaccinium macrocarpon 3 14.3 Equisetum fluviatile 2 14.3 Salix pedicellaris 23 14.3

(Russ.) C. Jens., and S. jensenii H. Lindb. (Vitt and Chee 1990). Although members of the Amblystegiaceae tend to prefer more minerotrophic microhabitats, W. fluitans and C. stramineum seem to occur in acid to intermediate acid habitats (ca. pH 3.7-6.0; Gorham and Janssens 1992b).

At Little Dollar Lake, several emergent vascular species col- onized this rich muck including Carex lasiocarpa, Dulichium arundinaceum (L.) Britton, Glyceria borealis (Nash) Batchelder, Juncus alpinus Vill., Potentilla palustris, and Puccinellia pallida (Torr.) R. T. Clausen. The only locality of Carex chordorrhiza L. f. was in the exposed peat of the southeastern lagg. Carex chor- dorrhiza is considered a species indicative of poor fens in Min- nesota (Wheeler et al. 1983) and rich fen conditions in Alberta (Vitt and Chee 1990).

1999] Hellquist and Crow—Little Dollar Lake Peatland 65

Table 5. Mean percent cover and percent frequency of dominant and as- sociated species in the Chamaedaphne calyculata cover type (220 quadrats). Species having less than 10% frequency are not included in the table.

*TWINSPAN indicator species.

Mean % Cover % Frequency BRYOPHYTES Sphagnum recurvum 74 83.2 Sphagnum majus 50 36.3 Sphagnum magellanicum* 14 26.5 Aulocomnium palustre 6 1S Sphagnum capillifolium* 18 10.6 VASCULAR SPECIES Chamaedaphne calyculata 46 99.1 Carex oligospe 26 94.7 Kalmia polifoli 15 62.8 Andromeda glaucophylla 11 36.3 cer rubrum <l 27.4 Vaccinium oxycoccos* 7 14.2

2. Chamaedaphne calyculata Cover Type

The maedaphne _ calyculata CT was unquestionably the most prominent cover type within the peatland (Figure 3). Mem- bers of the Ericaceae lent this cover type its scrubby, homoge- neous appearance. These shrubs occurred more or less continu- ously across a mosaic of hummocks and hollows carpeted by several species of Sphagnum. The Sphagnum recurvum—Carex oligosperma PHS was characterized by the grounded hummock- hollow complex. The Sphagnum magellanicum—Sarracenia pur- purea PHS formed narrow bands of quaking mat that fringed the majority of the lake (Figure 3).

The three dominant ericaceous shrubs of this cover type were Chamaedaphne calyculata, Kalmia polifolia Wangenh., and An- dromeda glaucophylla Link (Table 5). The prostrate ericad Vac- cinium oxycoccos was also apparent in this community, with in- dividuals winding over Sphagnum and between branches of eri- caceous shrubs on relatively open hummocks. Carex si

a, the most ubiquitous sedge of the peatland, was characteristic of this cover type

In northern Michigan, the Chamaedaphne calyculata associa- tion is the most ubiquitous in the region and is found extensively in almost every peatland (Gates 1942). Andromeda glaucophylla,

66 Rhodora [Vol. 101

Kalmia polifolia, Ledum groenlandicum Oeder, and Vaccinium oxycoccos are considered secondary components of the associa- tion (Gates 1942). All of the heath species at Little Dollar Lake, with the exception of L. groenlandicum, were readily apparent in the C. calyculata CT. At Little Dollar Lake, L. groenlandicum was scattered widely across the ericaceous scrub and was restrict- ed to large, dry hummocks.

Schwintzer (1981) noted that ‘‘bogs’’ (i.e. poor fens) in north- ern Michigan were often dominated by ericaceous shrubs and Sphagnum spp. and suggested that their dominance in these hab- itats was due to reduced rheotrophic conditions in these peatlands. Vitt and Slack (1975) found that Chamaedaphne calyculata had wide habitat preferences and therefore was not directly linked to any discrete association of Sphagnum species. In the Northeast, the C. calyculata association is affiliated with oligotrophic to om- brotrophic sites under very wet conditions (Damman and French 1987). The dominance of C. calyculata in large expanses of peat- land has been noted by many investigators (e.g. Crow 1969; Dun- lop 1987; Fahey 1993; Fahey and Crow 1995; Schwintzer 1981; Vitt and Bayley 1984; Vitt and Slack 1975).

e Chamaedaphne calyculata CT at Little Dollar Lake resem- bled the ‘“‘closed mat zone”’ of Vitt and Slack (1975) that was present in all eight of their northern Michigan study sites. This zone was distinguished by the presence of a tree layer of varying extent as well as an extensively developed ericaceous shrub layer (Vitt and Slack 1975). There was no tree canopy or evergreen

arkland at Little Dollar Lake, despite the presence of several coniferous tree species including Picea mariana (Miller) B.S.P, Abies balsamea (L.) Miller, Pinus strobus L., and Larix laricina (Duroi) K. Koch that were widely dispersed on the peatland mat. Picea mariana and L. laricina become abundant in peatlands with low water tables and well-drained peats (Glaser 1987). The re- duced prominence of these species at Little Dollar Lake may be indicative of a high water table.

€ most abundant conifer in this cover type was Pinus stro- bus, represented by both saplings and several mature trees grow- ing on the mat. Although P. strobus is not regarded as a wetland tree, this species is often found within peatlands (e.g. Crow 1969; Dunlop 1987; Fahey and Crow 1995; Miller 1996: Schwintzer 1981; C. E. Hellquist, ms. in prep.). In peatlands, P. strobus often grows to mature heights of several meters, but these individuals

1999] Hellquist and Crow—Little Dollar Lake Peatland 67

typically are unhealthy and chlorotic (Miller 1996; C. E. Hell- quist, pers. obs.).

2A. Sphagnum recurvum—Carex oligosperma Phase

This phase occupied expanses of grounded mat that exhibited the undulating topography characterized by Sphagnum hummocks that rise up to a meter above shallow, trough-like hollows (Figure 3). The species composition of the hummock-hollow complex is maintained in part by the growth and autogenic successional trends of Sphagnum species. Sphagnum grows apically over in- dividuals so that their stems build up a microtopography of un- dulating mounds (hummocks) that are supported by a scaffolding formed by the roots and branches of vascular plant species, es- pecially ericaceous shrubs (van Breeman 1995; Vitt et al. 1975).

A highly overlapping, directional succession of Sphagnum taxa occurs along the hummock-hollow gradient (Andrus 1986; drus et al. 1983; Crum 1988; Horton et al. 1979; Vitt and Slack 1984; Vitt et al. 1975). Sphagnum species characteristic of hum- mocks have high water-storing capacity, greater capillary pull, higher productivity under nutrient-deficient conditions, a greater cation exchange capacity due to the higher quantities of polyu-

1982: Moore and Bellamy 1974; Rydin 1985; Spearing 1972; van Breeman 1995; Vitt et al. 1975).

Due to the nutrient-poor nature of this cover type, the vege- tation was dominated by Sphagnum species including S. majus, S. recurvum, S. magellanicum Brid., and S. capillifolium (Ehrh.) Hedw. Ericaceous species including Chamaedaphne calyculata, Kalmia polifolia, Andromeda glaucophylla, and Vaccinium oxy- coccos, as well as Carex oligosperma were the other ubiquitous species of this cover type (Table 6). The four Sphagnum species are most abundant at the acid end of the pH spectrum from ap- proximately pH 3.7 to 5.0 (Gorham and Janssens 1992b). The prominence of C. calyculata, C. oligosperma, and Sphagnum spp. in open oligotrophic mats is a frequent association in northern Michigan (Schwintzer 1981).

Hollows are depressed areas where the water table typically pools at or just below the surface, often forming a waterlogged trough. These niches tend to have more plentiful nutrient supplies and higher pH values than hummocks (Crow 1969; Crum 1988;

68 Rhodora [Vol. 101

Table 6. Mean percent cover and percent frequency of dominant and as- sociated species in the Sphagnum recurvum—Carex oligosperma phase of the Chamaedaphne calyculata cover type (186 quadrats). Species having less than 10% frequency are not included in the table. *TWINSPAN indicator

Mean % Cover % Frequency

BRYOPHYTES Sphagnum recurvum* 58 85.5 Sphagnum magellanicum 34 53.8 Sphagnum capillifolium 22 29.0 Sphagnum majus* 47 25.8 Aulocomnium palustre 4 16.7 VASCULAR SPECIES Chamaedaphne calyculata 42 98.8 Carex oligosperma 22 93.5 Kalmia polifolia 17 TAS Andromeda glaucophylla* IS 43.0 Vaccinium oxycoccos* 25 39.2 r m ml 25.5

Moore and Bellamy 1974; Vitt and Slack 1975; Vitt et al. 1975). Sphagnum species that inhabit hollows have a loose, flimsy ap-

Klingrr. Sphagnum majus inhabited the lowest, dampest troughs of the

1999] Hellquist and Crow—Little Dollar Lake Peatland 69

scrub, and is a species that may grow submerged or emergent. It is characteristic of hollows in open sedge mats and low areas in open laggs (Crum 1983). Frequently found with S. majus in sat- urated hollows was the moss Warnstorfia fluitans, and two liv- erworts, the relatively abundant Cladopodiella fluitans (Nees) Joerg., and the more scarce and minute Cephaloziella elachista (Jack) Schiffn. Cladopodiella fluitans is typically found in sunken microhabitats (e.g. deer trails) where water may accumulate (Crum 1988). At Little Dollar Lake, both of these leafy liverwort species grew in similar microhabitats and were primarily found interwoven among moist Sphagnum stems.

Sphagnum recurvum was abundant along the upper edges of the hollows. Colonies of S. recurvum usually blended into S. ma- gellanicum and S. capillifolium along the sides of hummocks. Sphagnum recurvum is known to form “loose” carpets in the Great Lakes region (Crum 1983, 1988; Vitt et al. 1975). Sphag- num recurvum is abundant in mesotrophic microhabitats such as hollows in open peatland mats (Crum 1983), and thrives under acidic conditions with low calcium and magnesium concentra- tions (Vitt and Slack 1975). Sphagnum recurvum also has been noted as the dominant peat moss in some Ohio peatlands (An- dreas and Bryan 1990). In New York, S. recurvum is considered to be indicative of ‘“‘weakly minerotrophic” conditions and is found in poor fen habitats (Andrus 1980).

Growing among Sphagnum majus and S. recurvum was Carex oligosperma, the most abundant sedge in the peatland. Carex oli- gosperma inhabited open hollows throughout the grounded mat and is abundant in acidic northern Michigan basin peatlands (Vitt and Slack 1975). The presence of Chamaedaphne calyculata and C. oligosperma has been associated with oligotrophic nutrient re- gimes (Vitt and Bayley 1984). In the Red Lake peatland of Min- nesota, heliophilous C. oligosperma is a dominant species of open ombrotrophic and poor fen habitats within boreal patterned peat- lands (Glaser 1987; Wheeler et al. 1983). Common species that occur with C. oligosperma at Red Lake peatland are identical to species found at Little Dollar Lake including Andromeda glau- cophylla, C. calyculata, Eriophorum vaginatum L. [E. spissum Fern.], Kalmia polifolia, Ledum groenlandicum, and Vaccinium oxycoccos (Wheeler et al. 1983). In Maine, similar communities dominated by S. recurvum, S. magellanicum, C. oligosperma, and

70 Rhodora [Vol. 101

C. calyculata have been delineated over wide ranges of minero- trophic conditions (Anderson and Davis 1997).

At Little Dollar Lake, as elevation along the microtopograph- ical gradient increased, Sphagnum recurvum abundance dwindled and S. magellanicum became prominent. This same pattern has been observed in other northern Michigan peatlands (Vitt and Slack 1975; Vitt et al. 1975). In some areas of Little Dollar Lake peatland, S. papillosum had a patchy distribution among S. ma- gellanicum on low hummocks or on the sides of taller hummocks. Warnstorfia fluitans, a widespread northern moss species often found in acidic to moderately acidic conditions (Crum 1983; Jans- sens and Glaser 1986), also tended to inhabit moist nooks on the sides or bases of hummocks.

Near the tops of hummocks Sphagnum magellanicum faded out of prominence and blended into compact colonies of S. capilli- folium. Sphagnum magellanicum and S. capillifolium are typical of poor fen (oligotrophic) mat habitats (Andrus 1980; Crum 1983, 1988). Sphagnum magellanicum and S. capillifolium are known to initiate hummocks and grow on the sides or tops of low hum- mocks (Crum 1983, 1988; Vitt and Slack 1975). Sphagnum ma- gellanicum has an especially wide ecological amplitude across the hummock-hollow complex (Vitt and Slack 1975; Vitt et al. 1975).

Hummock tops are generally considered the driest, most nu- trient depleted, and most acidic microhabitats along the hum- mock-hollow sequence (Andrus 1986; Crum 1988: Vitt et al. 1975). The tallest hummocks at Little Dollar Lake often were crowned with compact populations of Sphagnum fuscum. Sphag- num fuscum is frequently found on hummock tops (Vitt et al. 1975). Sphagnum fuscum is typical of open acid peatland habitats and is strongly indicative of oligotrophic conditions (Crum 1983) as well as “‘ombrotrophic to weakly minerotrophic”’ conditions (Andrus 1986). Often intermingled with S. fuscum on larger hum- mocks were Polytrichum strictum Brid. [P. juniperinum Hedw. var. affine (Funck) Brid.], Calliergon stramineum, Dicranum un- dulatum Brid., Aulocomnium palustre (Hedw.) Schwaegr., and Bryum capillare Hedw. Polytrichum strictum is associated with dry, oligotrophic hummock tops (Andrus et al. 1983; Crum 1983: Janssens and Glaser 1986; Vitt 1990; Vitt et al. 1975). Calliergon stramineum prefers damp microhabitats in bogs and fens (Crum 1983) and can inhabit acidic to moderately acidic conditions

1999] Hellquist and Crow—Little Dollar Lake Peatland pe

(Gorham and Janssens 1992b). Dicranum undulatum is typically found in open peatland habitats especially on hummocks (Crum 1983).

An additional factor influencing the vegetation dynamics in this cover type was the presence of the caterpillars of the Chain-Spot- ted Geometer (Cingilia catenaria Drury, Lepidoptera: Geometri- dae), a pale white-colored moth whose caterpillars fed voracious- ly on ericaceous shrubs on the southwestern and southern mats of Little Dollar Lake peatland. This caterpillar is known to be an occasional pest on blueberry crops and has infested Ontario peat- lands, often with severe consequences (McGuffin 1987; Reader 1979). Despite the wide-ranging feeding preferences reported for the Chain-Spotted Geometer (Franklin 1948; McGuffin 1987), at Little Dollar Lake it was observed specifically defoliating Cha- maedaphne calyculata, Kalmia polifolia, and Andromeda glau- cophylla (see Hellquist 1996 for further details of the Little Dollar Lake infestations).

2B. Sphagnum magellanicum—Sarracenia purpurea Phase

The Sphagnum magellanicum—Sarracenia purpurea PHS formed the relatively stable floating mat that was present within five to twenty meters of the lake margin on all mats except the eastern and western mats (Figure 3). This quaking mat was sit- uated between the lake margin and the grounded mat and was easily distinguished by level expanses of peat mosses (“‘Sphag- num lawns’’). The Sphagnum lawns at Little Dollar Lake were best developed along the lake margin on the northwestern, south- western, and southern mats (Figure 3). These lawns had a gently undulating topography that was formed primarily by carpets of S. magellanicum, S. capillifolium, and S. papillosum. The lawn itself had the consistency of a saturated sponge, and was covered by sprigs of ericaceous shrubs, especially Chamaedaphne caly- culata and Kalmia polifolia (Table 7). The height differential of the low sprig-like ericads of the Sphagnum lawn and waist-high shrubby ericads on the grounded mat was conspicuous.

Sarracenia purpurea L., Scheuchzeria palustris L., and Erio- phorum virginicum L. thrived on the Sphagnum lawn where these species obtained their greatest prominence. In Ontario, where Chamaedaphne calyculata was less prominent, oligotrophic in- dicator species such as Scheuchzeria palustris, Eriophorum spp. and bryophytes including Cladopodiella fluitans, Sphagnum ma-

qT Rhodora [Vol. 101

Table 7. Mean percent cover and percent frequency of dominant and as- sociated species in the Sphagnum magellanicum—Sarracenia purpurea phase

species.

Mean % Cover % Frequency BRYOPHYTES Sphagnum magellanicum 31 100.0 Sphagnum capillifolium 30 64.7 Sp majus* 28 58.8 Sphagnum papillosum 30 55.9 Sphagnum recurvum 20 38.2 Cladopodiella fluitans = 23.5 Sphagnum cuspidatum 18 a9 VASCULAR PLANTS hamaedaphne calyculata 23 100.0 Kalmia polifolia 14 91.2 Vaccinium oxycoccos 23 88.2 Carex o. perma 20 88.2 Andromeda glaucophylla 8 67.6 Sarracenia purpurea* 10 52.9 Drosera rotundifolia 1 52.9 Eri rum virginicum 2 26.5 Rhynchospora alba 3 14.7

Jus, S. recurvum, and Warnstorfia fluitans, had an increased prev- alence (Vitt and Bayley 1984). At Little Dollar Lake, this same suite of species was observed on the Sphagnum lawns.

The Sphagnum lawn was also the only habitat in the peatland where the orchids Calopogon tuberosus (L.) B.S.P. and Pogonia ophioglossoides (L.) Ker Gawler grew. Calopogon tuberosus grew in all Sphagnum lawn habitats, whereas P. ophioglossoides was limited to the lawn of the northwestern mat. Both of these orchids are abundant in northern Michigan peatlands, but were surprisingly scarce at Little Dollar Lake.

The Sphagnum magellanicum—Sarracenia Purpurea PHS on the northwestern floating mat was pock-marked by large, irregular holes in the mat that often were colonized by two submerse species, Potamogeton confervoides Reichb. and Utricularia gem- iniscapa Benj. Presumably, these holes had been accentuated and possibly maintained by beaver activity. The edges of these holes were fringed with exposed peat and were the only sites for Ly-

1999] Hellquist and Crow—Little Dollar Lake Peatland 73

copodiella inundata (L.) Holub [Lycopodium inundatum L.] and Drosera intermedia Hayne. Other species that grew on this ex- posed peat included D. rotundifolia L., Eriocaulon aquaticum (Hill) Druce [E. septangulare With.], Menyanthes trifoliata L., Rhynchospora alba (L.) Vahl., and the liverwort Cladopodiella fluitans.

The Sphagnum magellanicum—Sarracenia purpurea PHS closely resembled a similar community at Mud Pond Bog (Moul- tonborough, NH). Both lawn communities were saturated, spong habitats extensively colonized by insectivorous plants (Sarracenia purpurea, Drosera intermedia, and D. rotundifolia), as well as stunted ericaceous species, especially Chamaedaphne calyculata (C. E. Hellquist, ms. in prep.). These lawn habitats at Moulton- borough were also the preferred habitats of Pogonia ophioglos- soides and Calopogon tuberosus (C. E. Hellquist, ms. in prep.) and corresponded well to the Sphagnum lawn concept of Crum (1988). The Vaccinium oxycoccos—Rhynchospora alba subtype of Mud Pond Bog (Hillsborough, NH), characterized by dwarfed ericaceous shrubs as well as D. rotundifolia and S. purpurea (Dunlop 1987), also was similar to the species assemblage at Little Dollar Lake. In Ohio, similar species inventories have

een recorded in Sphagnum lawn communities (Andreas and Bryan 1990).

3. Chamaedaphne calyculata—-Triadenum fraseri Cover Type

e Chamaedaphne calyculata—Triadenum fraseri CT was composed of two phases that exhibited possible terrestrialization patterns in two distinct areas of the peatland. One phase of this cover type was associated with terrestrialization at the lake mar- gin (the Sphagnum papillosum PHS) and the other phase was associated with terrestrialization of the stream channel (the S. majus PHS; Figure 3). This cover type has been renamed and is identical to the C. calyculata—Carex lasiocarpa CT first described by Hellquist (1996) at Little Dollar Lake.

This cover type represents a hybrid association colonized by species common to both the Chamaedaphne calyculata CT and the Calamagrostis canadensis CT (Table 8). Few species were restricted to this cover type, thus it is defined more by the absence of certain species than those present in its marsh-like physiog- nomy. Despite its low cover value, Triadenum fraseri (Spach) Gleason was a frequent component of this cover type and was

74 Rhodora [Vol. 101

Table 8. Mean percent cover and percent frequency of dominant and as- sociated species in the Chamaedaphne calyculata—Triadenum fraseri cover type (32 quadrats). Species having less than 10% frequency are not included in the table. *TWINSPAN indicator species. Sphagnum cuspidatum also was designated as an indicator species, but does not appear in the table.

Mean % Cover % Frequency

BRYOPHYTES Sphagnum recurvum* 45 50.0 Sphagnum majus* 60 34.4 Sphagnum papillosum* 44 28.1 Sphagnum magellanicum 13 18.8

VASCULAR PLANTS Chamaedaphne calyculata 44 93.8 Triadenum fraseri fa, 59.4 Carex lasiocarpa 31 56:3 Carex oligosperma* 15 34.4 Vaccinium macrocarpon 26 34.4 Carex ca - 25 25.0 Calamagrostis canadensis 24 21.9 Drosera rotundifolia 21.9

1 Carex utriculata 4 Glyceria canadensis 2 Acer rubrum <1 18.8 Potentilla palustris 3 Sarracenia purpurea* 4

readily apparent with its distinct growth form and reddish-hued foliage (Table 8). Triadenum fraseri was commonly seen in more minerotrophic microhabitats along the lakeshore, stream channel, and in the lagg.

Although prominent in the Calamagrostis canadensis CT, Car- ex lasiocarpa was also conspicuous in this cover type with a cover of 31% and 56.3% frequency. In the stream channel, the distinct physiognomy of the C. lasiocarpa community traced the course of the stream channel from its origin in the muck pools of the southern mat extension, through the ericaceous scrub of the southern mat to the open water of the lake (Figures 1 and 3).

Carex lasiocarpa is indicative of more minerotrophic condi- tions (e.g., Crum 1988; Jeglum 1971; Schwintzer 1978; Vitt and Bayley 1984; Vitt and Slack 1975). With its prolific interlaced rhizome sytems, C. lasiocarpa is one of the most important lake- fill species that can initiate the process of terrestrialization along shorelines (Gates 1942). In northern Michigan, C. lasiocarpa is

1999] Hellquist and Crow—Little Dollar Lake Peatland 75

Table 9. Mean percent cover and percent frequency of dominant and as- sociated species in the Sphagnum majus phase of the Chamaedaphne caly- culata—Triadenum fraseri cover type (22 quadrats). Species having less than 10% frequency are not included in the table. TWINSPAN indicator species.

Mean % Cover % Frequency

BRYOPHYTES Sph eb recurvum* 51 63.6 Sphagnum majus* 64 45.5

VASCULAR PLANTS Chamaedaphne calyculata 43 95.5 Carex oligosperma* 15 45.5 Calamagrostis canadensis* 28 219 Vaccinium macrocarpon* 40 25S Carex utricu 5 0 Ue | Glyceria canadensis 3 22-0 Acer rubrum <1 Fb ied Potentilla palustris 5 18.2 Drosera dunia <1 18.2 Carex canescen 8 13.6

the primary mat-forming species along more alkaline lake mar- gins with circumneutral pH values (Crum 1988; Vitt and Slack 1975). A circumneutral sedge mat dominated by C. lasiocarpa has also been noted in southern Michigan (Crow 1969). At Red Lake peatland in Minnesota, C. lasiocarpa was the most abundant vascular plant in open rich fens (Wheeler et al. 1983). Similarly, in north-central New Hampshire, the presence of C. /asiocarpa along lake edges and in wet habitats with a fen-like flora has also been noted (Fahey and Crow 1995; C. E. Hellquist, ms. in prep.).

3A. Sphagnum majus Phase

The Sphagnum majus PHS was a relatively minerotrophic phase that was contained within the stream channel that origi- nated in the southeastern lagg of the southern mat extension, and eventually reached the open water of Little Dollar Lake. The S. majus PHS also comprised the narrow eastern and western mats (Figure 3). Both of these mats bordered steep upland slopes and were dominated by S. majus and S. recurvum (Table 9). Con- spicuous vascular plants on these mats included Carex canescens, C. lasiocarpa, Calla palustris L., Chamaedaphne calyculata, Vaccinium macrocarpon Aiton, and Triadenum fraseri. The rel- ative minerotrophy of this vegetation phase was strongly sug-

76 Rhodora [Vol. 101

gested by the presence of S. subsecundum. This minerotrophic species is often found in open, wet lagg habitats and sedge mats (Andrus 1980; Crum 1983, 1988), and has been shown to grow in moderately acidic habitats (ca. pH 4.5—6.0; Gorham and Jans- sens 1992b). This phase was the only area in the peatland where S. subsecundum grew in abundance (Hellquist and Crow 1997).

3B. Sphagnum papillosum Phase

e Sphagnum papillosum PHS occurred along the majority of the lake margin. This community was distinguished by a narrow swath (~1.0 m) of Chamaedaphne calyculata and Kalmia poli- folia intermingled with Vaccinium macrocarpon and Triadenum fraseri that extended into the open water of the lake. Immediately behind this band of C. calyculata was a wet trough with some standing water, colonized primarily by S. papillosum, Carex ca- nescens L., C. limosa L., and Rhynchospora alba. Although sam- pled infrequently, C. limosa was a major component of this phase. Other species frequent in this trough included Cladopodiella flui- tans, Sphagnum cuspidatum, S. recurvum, Carex lasiocarpa, and Sarracenia purpurea (Table 10).

The Sphagnum papillosum PHS of Little Dollar Lake closely resembled an ‘‘acid lake edge” community (Crum 1988; Vitt and Slack 1975) that was also observed at Heath Pond Bog, New Hampshire (Fahey 1993). The acid lake vegetation sequence is characterized by the lack of vascular macrophytes in the lake itself, the presence of Chamaedaphne calyculata growing out into open water, Sphagnum as a pioneer among C. calyculata, fre- quently a “sparse Carex fringe”’ along the lake, and a narrow Sphagnum lawn that quickly merges into a grounded mat (Crum 1988).

Vitt and Slack (1975) defined the “acid lake edge” community by the presence of a Sphagnum cuspidatum—Sphagnum papillos- um—Carex limosa—Carex paupercula association and a pH range

tats (Vitt and Slack 1975), although along the lake margin at Little Dollar Lake (mean pH = 4.5, n = 40) S. papillosum grew both emergent above the water level, and partially submerged

1999] Hellquist and Crow—Little Dollar Lake Peatland 77

Table 10. Mean percent cover and percent frequency of dominant and associated species in the Sphagnum papillosum phase of the Chamaedaphne calyculata—Triadenum fraseri cover type (10 quadrats). Species having less than 20% frequency are not included in the table. *TWINSPAN indicator species.

Mean % Cover % Frequency

BRYOPHYTES Sphagnum papillosum 49 80.0 Sphagnum magellanicum 9 40.0 Sphagnum cuspidatum 41 40.0 Sphagnum recurvum 3 20.0 Sphagnum capillifolium 10 20.0 VASCULAR PLANTS Chamaedaphne calyculata 47 90.0 Triadenum fraseri - 80.0 Carex canescens 36 50.0 Vaccinium macrocarpon 10 50.0 Carex lasiocarpa* 22 40.0 Sarracenia purpure 4 40.0 Drosera rotundifolia 30.0 Eriocaulon aquaticum 2a 30.0 Kalmia polifolia 3 30.0 Vaccinium oxycoccos 1 30.0 Rhynchospora alba <l 20.0 Scheuchzeria palustris 5 20.0 Utricularia geminiscapa <1 20.0

among Chamaedaphne calyculata branches in the lake (Hellquist 1996). In northern Michigan, Sphagnum papillosum is known to in- habit wetter, more mineral rich habitats on mats or lake margins influenced by water movement (Crum 1983, 1988). In Ontario, S. papillosum is also associated with lake edge vegetation se- quences (Vitt and Bayley 1984). The liverwort Cladopodiella fluitans also seems to indicate the acid nature of the lake margin community based on its preferred pH range of 3.7 to 5.0 (Gorham and Janssens 1992b). In Maine, sedge-moss lawn communities characterized by S. papillosum, S. magellanicum, Carex limosa, Rhynchospora alba, and Scheuchzeria palustris were found in ‘‘very acidic” to “‘moderately acidic” fens (Anderson and Davis

‘ Vitt and Slack (1975) cite Carex limosa as being essentially restricted to this zone. At Little Dollar Lake, C. limosa was pri-

78 Rhodora [Vol. 101

marily found along the lake margin with the exception of a few scattered localities in wet areas of the Calamagrostis canadensis CT. Of the four species described as characteristic of this com- munity by Vitt and Slack (1975), only Carex paupercula Michx. was absent at Little Dollar Lake.

4. Potamogeton confervoides—Utricularia geminiscapa Cover Type

The depauperate aquatic flora of Little Dollar Lake reflected the acidity of the lake water. The aquatic flora consisted of three submersed species (Jsoétes echinospora Durieu, Utricularia gem- iniscapa, and Potamogeton confervoides) and two floating species (Nuphar variegata Durand and Nymphaea odorata Aiton). Two submersed and/or emergent species, Eriocaulon aquaticum and Hypericum boreale (Britton) E. Bickn. forma callitrichoides Fas- sett, were also present in or along the lake (Hellquist 1996; Hell- quist and Crow 1997).

Potamogeton confervoides and Utricularia geminiscapa were concentrated along the periphery of the floating mat, among the submerged rhizomes and roots of ericaceous shrubs growing into the open water of the lake. While P. confervoides was especially abundant and fruited copiously in the summer of 1994, it seemed less abundant in 1995, and was not fruiting as prolifically as in 1994. In 1996 the population continued to recede, with the ma- jority of the population scattered along the lake margin of the eastern and southern mats. Previously, it had been found around the entire circumference of the lake

In Michigan, Potamogeton confervoides occurs locally in lakes and acid bogs (Voss 1972) and is listed as a state of Michigan Threatened Species (Anonymous 1994; Beaman et al. 1985), al- though its status as Threatened may be due to under-collection rather than to actual rarity. In New England, P. confervoides is associated with soft waters, including high elevation ponds and lakes (Hellquist 1980). It is characteristic of extremely acidic wa- ters with low alkalinity values (maximum alkalinity value 8.5 mg/l CaCO,) and is the only pondweed found in Sphagnum bog ponds (Hellquist 1980).

Utricularia geminiscapa was also found submersed along the edge of the floating mat. In New England, U. geminiscapa is associated with bog ponds and acidic to moderately alkaline wa- ters with a pH of 3.5-8.6 and alkalinity of 5.4-69.5 mg/l CaCO,

1999] Hellquist and Crow—Little Dollar Lake Peatland 79

(Crow and Hellquist 1985). Like Potamogeton confervoides, the abundance of U. geminiscapa was noticably lower in 1996 than in the two previous years.

Isoétes echinospora grew submersed in the sandy sediments on the bottom of Little Dollar Lake. Individuals tended to grow widely scattered and were most conspicuous in shallow water off the eastern mat. In 1996, Hypericum boreale forma callitrichoi- des, an aquatic to partially emergent growth form, was thriving in the water along the edge of the eastern mat.

Evidence of terrestrialization and small-scale paludification processes at Little Dollar Lake. In an analysis of the pollen and sediment stratigraphy of Little Dollar Lake, Futyma (1982) stated that since the formation of the peatland approximately 3500 years before present, at least fifty percent of the surface area of Little Dollar Lake had been covered by peatland vegetation. At the time of this study, evidence of these terrestrialization patterns was visible at Little Dollar Lake in the former stream channel as well as in the muck pool area of the southern mat extension.

Apparent botanical evidence of terrestrialization was observed in the muck pools of the Sphagnum cuspidatum—Dulichium arun- dinaceum PHS of the Calamagrostis canadensis CT, and in the various areas of the S. majus and S. papillosum phases of the Chamaedaphne calyculata-Triadenum fraseri CT (Hellquist 1996). In these areas there was evidence from relict aquatic veg- etation, as well as from aerial photography, that previously aquat- ic habitats were being colonized by emergent wetland vegetation (Hellquist 1996).

At Little Dollar Lake peatland, apparent on-going terrestriali- zation was most conspicuous at the muck pools in the southeast- ern lagg of the southern mat extension. These muck pools ex-

rienced conditions that vacillated between exposure and sub- mergence. Along the edges of these mucky areas Chamaedaphne calyculata, Carex lasiocarpa, Dulichium arundinaceum, and Gly- ceria canadensis, as well as other emergent vascular species, were well established and were apparently encroaching into the open areas of peaty muck.

In 1994 and 1995, the muck pools contained stranded aquatic vascular plant species that were directly subjected to the fluctu- ating water levels of the peatland. One of the stranded aquatic species was Potamogeton oakesianus Robbins, which grew stunt-

80 Rhodora [Vol. 101

ed and infertile with ‘floating’ leaves lying on the surface of the mucky peat. Another stranded aquatic species, Sparganium min- imum (Hartman) Fries, was found fruiting abundantly along the edges of the muck pools. Utricularia intermedia Hayne also grew in the peaty muck. In 1995, U. intermedia was extremely scarce with less than ten individuals located. However, when the muck pools were flooded in 1996, U. intermedia was frequent, with many lush, thriving individuals growing among the emergent veg- etation.

There was also evidence of terrestrialization in the stream chan- nel that wound through the southern mat and southern mat ex- tension. The intermittently aquatic stream channel was colonized by a tenuous, loose network of Carex lasiocarpa that readily sank and quaked extensively when walked upon. This channel was also home to approximately 25 emergent clusters of stranded Nuphar variegata that were surrounded entirely by C. lasiocarpa and scattered individuals of Potentilla palustris in an approximately 3.0 m X 3.0 m area (Hellquist 1996). In a 1978 infrared aerial photograph, this stream channel was visible as open water where- as 1995 aerial photos and concurrent ground surveys clearly showed that, although still visible, the channel had become en- tirely covered by a quaking mat of vegetation dominated by C. lasiocarpa (see Hellquist 1996 and Hellquist and Crow 1997 for photographs).

Water levels in 1994 and 1995 were relatively low, but in 1996 the stream channel and muck pool areas became partially sub- merged by high water levels. In most areas, the Sphagnum majus PHS was covered by water with a minimum depth of ca. 20 cm. Presumably during high water years, terrestrialization by emer- gent sedges and other wetland taxa such as Potentilla palustris probably does not progress as rapidly as in drier years. Therefore, if the peatland experiences more dry years than wet years, ex- pansion of the sedge-dominated mat would continue, and peat deposition and plant colonization would further promote terres- trialization. This process could occur relatively rapidly in an area as confined as the stream channel.

A similar rapid terrestrialization process has occurred at Weber Lake Bog in Cheboygan County, Michigan. A small pool of open water fringed by Carex lasiocarpa was mapped by University of Michigan Biological Station ecology classes in 1967 (Vitt and Slack 1975). When examined by Vitt and Slack in the mid-1970s

1999] Hellquist and Crow—Little Dollar Lake Peatland 81

neither the open water nor C. lasiocarpa was present. The pool was replaced by an open mat community with stranded clumps of Dulichium arundinaceum. During the summer of 1995, the area that had once been open water was still apparent, and con- tained stranded individuals of Nuphar variegata, as well as a few isolated individuals of C. limosa. Carex lasiocarpa has remained absent at Weber Lake Bog (E. G. Voss and C. E. Hellquist, pers. obs. 1995).

In some lagg areas of the Jris versicolor—Lycopus uniflorus PHS at Little Dollar Lake, small-scale paludification was ob- served where Sphagnum spp. were growing onto upland slopes. The process of paludification often is associated with rising water levels, especially in shallow, flat basins (Crum 1988). In some basin peatlands in northern Michigan, paludification is initiated by the compression of the lowermost peat layers. These com- pacted layers become so tightly condensed that they act as a seal- ant that prevents water from percolating out of the basin. Thus, any water entering the basin is retained by the more porous upper peat layers (Futyma 1982). This sequence is believed to have resulted in paludification at other Michigan peatlands, including Tahquamenon Bog and the Trout Lake peatlands in Chippewa County (Futyma 1982) and Lake Sixteen peatland in Cheboygan County (Futyma and Miller 1986). At Little Dollar Lake, the gradual creep of Sphagnum onto upland soils in some areas can probably be attributed to pooling of water trapped in the lagg.

Based on pollen and stratigraphic analysis from Futyma (1982), as well as field and photographic evidence described by Hellquist (1996) and Hellquist and Crow (1997) there seems to be satis- factory botanical evidence illustrating terrestrialization and pal- udification processes at Little Dollar Lake peatland. The patterns observed follow the ‘“‘bog climax model’’ of peatland succession proposed by Klinger (1996), where areas of water are colonized by extensive, but irregular growth of vegetation while paludifi- cation proceeds as peat is deposited over upland soils along the margin of the bog basin. At Little Dollar Lake the areas along the lakeshore, stream channel, and muck pools with stranded aquatic taxa seem to be indicative of terrestrialization processes whereas paludification is less easily discerned, but still present in

some lagg areas.

Little Dollar Lake peatland classification. The majority of

82 Rhodora [Vol. 101

northern Michigan peatlands exist in glacial topography and re- main under the influence of local hydrology. Despite the fact that most of these peatlands are dominated by Sphagnum, they are best classified as fens (Vitt et al. 1975). Little Dollar Lake cor- responds well to the delineation of a northern Michigan poor fen or “bog”’ (i.e. a more acid, oligotrophic peatland) as defined by Schwintzer (1981). Schwintzer (1981) states that these peatlands are weakly minerotrophic with low pH values (3.8—4.3) and low concentrations of calcium cations (1.2-3.7 mg/L). These poor fen complexes have a prominent Sphagnum cover, low vascular spe- cies richness, well developed open areas dominated by ericaceous shrubs, and reduced tree cover (Schwintzer 1981).

This study and its companion studies (Hellquist 1996; Hellquist and Crow 1997) have documented the flora and described the vegetation patterns within Little Dollar Lake peatland. This anal- ysis, used in conjunction with the postglacial history of the Little Dollar Lake basin (Futyma 1982), has laid a foundation for sub- sequent studies that may further elucidate the dynamic processes (e.g., hydrology, nutrient regimes, interspecific plant interactions, plant-herbivore interactions, successional patterns) that influence the abundance and distribution of peatland vegetation at Little Dollar Lake.

ACKNOWLEDGMENTS. The senior author wishes to thank to Dr Thomas Lee and Scott Miller for their patient assistance with TWINSPAN; Dr. Lee and Dr. Janet Sullivan also provided advice and editiorial suggestions on earlier drafts of this research; Dr. Edward Voss encouraged this project and has provided generous support throughout its duration from specimen annotations to comments on early thesis drafts; Dr. Howard Crum kindly as- sisted with bryophyte taxonomy; the efforts of Dean I. Reid and the staff of the Naubinway Field Office of the Michigan Depart- ment of Natural Resources are also appreciated; the technological expertise of Chris Cerrudo was essential to the completion of various graphics. The thoughtful editorial comments and sugges- tions of two anonymous reviewers were also appreciated. The financial and logistical support of both the University of New Hampshire Department of Plant Biology and the University of Michigan Biological Station (including a Henry Allan Gleason Fellowship to the senior author in 1995) is gratefully acknowl- edged. This study was conducted in partial fulfillment of the re-

1999] Hellquist and Crow—Little Dollar Lake Peatland 83

quirements for the Master of Science degree in Plant Biology for the University of New Hampshire, Durham, NH. This paper is Scientific Contribution No. 1984 from the New Hampshire Ag- ricultural Experiment Station.

LITERATURE CITED

ANDERSON, D. S. AND R. B. Davis. 1997. The vegetation and its environments in Maine peatlands. Canad. J. Bot. 75: 1785-1805.

ANDREAS, B. K. AND G. R. BRYAN. 1990. The vegetation of three Sphagnum dominated basin-type bogs in northeastern Ohio. Ohio J. Sci. 90: 54—

66. ANnprus, R. E. 1980. Sphagnaceae (Peat Moss Family) of New York State. Contributions to a Flora of New York State III. N.Y. State Museum Bull. o. 422

. 1986. Some aspects of Sphagnum ecology. Canad. J. Bot. 64: 416— 426.

. J. WAGNER, AND J. E. Titus. 1983. Vertical zonation of Sphagnum

concern, and probably extirpated. Endangered Species Program of the Michigan Department of Natural Resources and the Michigan Natural Features pete sine MI.

BEAMAN, J. oe A. Bourpo, FE W. Case, S. R. Crispin, D. HENSON, R. — Pippen, A. A. aheacvonts E. G. Voss, AND P. W. THOMPSON. 1985. dangered and threatened vascular plants in Michigan. II. Third aces review proposed list. Mich. Bot. 24: 99-116

Criymo, R. S. 1963. Ion-exchange in Sphagnum and its relation to bog ecol- ogy. Ann. Bot., n.s. 27: 309-324.

. 1964. The origin of acidity in Sphagnum bogs. Bryologist 67: 427—

431. AND P. M. Haywarp. 1982. The ecology of Sphagnum, pp. 229-289. In: = J. E. Smith, ed., Bryophyte Ecology. Chapman and Hall, New

Yor!

Crow, = E. 1969. An ecological analysis of a southern Michigan bog. Mich. Bot cn 11-27.

pC. B. HELLQutst. 1985. Aquatic Vascular Plants of New England. bak 2. Lentibulariaceae. New Hampshire Agric. Exp. Sta. Bull. 528.

Crum, H. A. 1983. Mosses of the Great Lakes Forest, 3rd ed. Univ. Michigan Herbarium, Ann Arbor, MI.

_ 1988. A Focus on Peatlands and Peat Mosses. Univ. Michigan Press, Ann Arbor, MI.

Daman, A. W. H. AND T. W. FRENCH. 1987. The ecology of peat bogs of the glaciated northeastern United States: A community profile. U.S. Fish Wildl. Serv. Biol. Rep. 85(7.

DANSEREAU, P. AND F SEGADAS- VIANNA. 1952. Ecological study of peat bogs of eastern North America. I. Structure and evolution of vegetation. Can- ad. J. Bot. 30: 490-520

84 Rhodora [Vol. 101

ne D. A. 1987. Community classification of the vascular vegetation of w Hampshire peatland. Rhodora 89: 415—440.

Fas, L. L. 1993. A vegetation, floristic, and phytogeographic analysis of

o New Hampshire peatlands. M.S. thesis, Univ. New Hampshire, Dur- ey NH.

AND G. E. Crow. 1995. The vegetation of Pequawket Bog, Ossipee,

New Hampshire. Rhodora 97: 39-92.

FRANKLIN, H. J. 1948. Cranberry Insects in Massachusetts. Massachusetts

a:

ic. Exp. Sta. Bull. FutyMa, R. P. 1982. Postglacial vegetation of eastern Upper Michigan. Ph.D. pre Univ. Michigan, Ann Arbor, MI.

N. G. MILLER. 1986. Stratigraphy and genesis of the Lake Six- nt aeuieak northern Michigan. Canad. J. Bot. 69: 3008-3019. Gates, F C. 1942. The bogs of northern lower Mickdeans Ecol. Monogr. 12: Gaucu, H. G. 1982. Multivariate Analysis in Community Ecology. Cam- bridge Univ. Press, New York. Gicnac, L. D. AND D. H. Vitr. 1994. Responses of northern peatlands to climate change: Effects on bryophytes. J. Hattori Bot. Lab. 75: 119-132. GLASER, P. H. 1987. The ecology of patterned boreal peatlands of northern Minnesota: A community profile. U.S. Fish Wildl. Serv. Biol. Rep.

Gore, A. J : P 1983. Introduction, pp. 1-34. In: A. J. P. Gore, ed., Ecosystems of the World. 4A. Mires: Swamp, Bog, Fen, and Moor. Elsevier, Am-

sterdam. GorHAM, E. 1957. The development of peatlands. Quart. Rev. Biol. 32: 145— 166

. 1961. Factors influencing supply of major ions to inland waters, with i references to the atmosphere. Bull. Geol. Soc. Amer. 72: 795—

ate Biotic impoverishment in northern peatlands, pp. 65-98. J

G. M. Woodwell, ed., The Earth in Transition: Patterns and Processes Biotic Impoverishment. Cambridge Univ. Press, New York.

. A. JANSSENS. 1992a. The paleorecord of geochemistry and hydrology of northern peatlands and its relation to global change. Suo 43: fica 126.

———.. 1992b. Concepts of fen and bog re-examined in relation to myopic cover and the acidity of surface waters. ACTA Soc. Bot. Poloniae 61: 7—20. HELLQuisT, C. B. 1980. Correlation of alkalinity and the distribution of Po- tamogeton in New England. Rhodora 82: 331-344. pe C. E. 1996. A flora and analysis of the bryophyte and vascular communities of Little Dollar Lake peatland, Mackinac County, pesiheas M.S. thesis, Univ. New Hampshire, Durham, NH. AND G. E. Crow. 1997. The bryophyte and vascular flora of Little Dollar Lake peatland, Mackinac County, Michigan. Rhodora 99: 195— 222.

HEMoND, H. F 1980. Biogeochemistry of Thoreau’s bog, Concord, Massa- chusetts. Ecol. Monogr. 50: 507-526.

1999] Hellquist and Crow—Little Dollar Lake Peatland 85

Hitt, M. O. 1979. TWINSPAN—A FORTRAN program for arranging mul- tivariate data in an ordered two-way table by classification of the indi- viduals and attributes. Ecology and Systematics, Cornell Univ., Ithaca,

Horton, D. G., D. H. Vitt, AND N. G. SLACK. 1979. Habitats of circumbo- ae Sphagna: I. A quantitative analysis and review of species in the Caribou Mountains, northern Alberta. Canad. J. Bot. 57: 2283-— 23

eS i J. A. AND P. H. GLASER. 1986. The bryophyte flora and major peat rming mosses at Red Lake peatland, Minnesota. Canad. J. Bot. 64: —442. JeEcLuM, J. K. 1971. Plant indicators of pH and nese o in peatlands at Candle Lake, Saskatchewan. Canad. J. Bot. 49: —1676. KILHAM, P. 1982. The biogeochemistry of bog hone and the chemical ecology of Sphagnum. Mich. Bot. 21: 159-168. Kurncer, L. E 1996. The myth of ~ classic hydrosere model of bog suc- cession. opi Alpine Res. 28: McGurrin, W. C. 1987. Guide to “he Geciieisides of Canada ca i aie II. Subfamily once 4. Mem. Entom. Soc. Canada, no. 138. Miter, S. D. 1996. A vegetation and dendroecological slates of Sa le Bog, Durham, NH. M.S. thesis, Univ. New Hampshire, Durham

Moore, P. D. AND D. J. BELLAMY. 1974. Peatlands. Springer-Verlag, New

York. PieLou, E. C. 1979. Biogeography. John Wiley and Sons, New York. READER, R. 1979. Impact of leaf-feeding insects on three bog ericads. Canad. J. Bot. 57: 2107-2112 Rice, S. K. AND P. H. Scueupp. 1995. On the ecological and evolutionary significance of branch and leaf morphology in aquatic Sphagnum 3-846.

(Sphagnaceae). Amer. J. Bot. 82: 833 Risser, P. G. 1995. The status of the science examining ecotones. BioScience 4S: 318- 325.

Rypin, H. 1985. Effect of water level on dessication of Sphagnum in relation to surrounding Sphagna. Oikos 45: 374-379.

SCHWINTZER, C. R. 1978. Vegetation and nutrient status of northern Michigan fens. Canad. J. Bot. 56: 3044-3051.

_ 1981. Vegetation and nutrient status of northern Michigan bogs and

conifer swamps with a comparison to fens. Canad. J. Bot. 59: 842-853.

p G. WILLIAMS. 1974. Vegetation changes in a small Michigan bog aha 1917 to 1972. Amer. Midl. Nat. 92: 447-459

SHIMWELL, D. W. 1971. The Description and Classification of Vegetation.

niv. ie

SLACK, N. G., D. H. Vitt, AND D. G. Horton. 1980. Vegetation gradients of

minerotrophically rich oo in western Alberta. Canad. J. Bot. 58: 330— 350.

SPEARING, A. M. 1972. Cation-exchange capacity and galacturonic acid con- tent of several species of Sphagnum in Sandy Ridge Bog, Central New

York State. Bryologist 75: 154-158. TRANSEAU, E. N. 1903. On the geographic distribution and ecological rela-

86 Rhodora [Vol. 101

tions of the bog plant societies of northern North America. Bot. Gaz. 36: 401—420.

VAN BREEMAN, N. 1995. How Sphagnum bogs down other plants. Trends Ecol. Evol. 10: 270-275

VAN TONGERAAN, O. E R. 1987. Cluster sea = 174-212. In: R. H. G.

ongman, C. J. E ter Braak, and O. FE R. van Tongeren, eds., Data Anal- ysis in Community and Landscape oe Pudoc, Wageningen, Neth- erlands.

Virt, D. H. 1990. Growth and production rages of boreal mosses over cli- matic, chemical and topographic gradients. J. Linn. Soc. Bot. 104: 35—59.

AND S. BAYLEY. 1984. The eae and water chemistry of four

oligotrophic basin mires in northwestern Ontario. Canad. J. Bot. 62:

1485-1500.

AND W.-L. CHEE. 1990. The relationship of vegetation to surface wa-

ter chemistry and peat chemistry in fens of Alberta, Canada. Vegetatio

89: 87-106.

» H. A. Crum, AND J. A. SNIDER. 1975. The vertical zonation of Sphag- num species in hummock-hollow complexes in northern Michigan. Mich. Bot. 14: 190-200.

———, D. G. Horton, N. G. SLACK, AND N. MALMER. 1990. Sphagnum- dominated peatlands of the hyperoceanic British Columbia coast: Pat- terns in surface water chemistry and vegetation. Canad. J. For. Res. 20: 696-711.

» Y. Li, AND R. J. BELLAND. 1995. Patterns of bryophyte diversity in peatlands = continental western Canada. Bryologist 98: 218—227.

N. G. Stack. 1975. An analysis of the vegetation of Sphagnum- phi kettle hole bogs in relation to environmental gradients. Canad. J. Bot. 53: 332-359.

AND . 1984. Niche diversification of Sphagnum relative to en- vironmental factors i in northern Minnesota peatlands. Canad. J. Bot. 62: 1409-1430.

Voss, E. S. 1972. Michigan Flora. Part I. Gymnosperms and Monocots. Cran- pati Institute of Science Bull. 55 and Univ. Michigan Herbarium, Ann

Ar aaa a e P. H. Graser, E. Goruam, C. M. WETMoRE, FE D. Bowers, Dib. «Ay Janssens 1983. Contributions to the flora of the Red Lake

a » northern Minnesota, with special attention to Carex. Amer. Midl. Natl. 110: ae

RHODORA, Vol. 101, No. 905, pp. 87-91, 1999

NOTE

MORE MOLECULAR EVIDENCE FOR INTERSPECIFIC RELATIONSHIPS IN LIQUIDAMBAR (HAMAMELIDACEAE)

JIANHUA LI AND MICHAEL J. DONOGHUE

Harvard University Herbaria, 22 Divinity Ave., Cambridge, MA 02138

There are four species of Liqguidambar L. (Altingioideae, Ha- mamelidaceae). Liguidambar orientalis Miller occurs in western Asia and L. styraciflua L. in Northern and Central America, while L. acalycina Chang and L. formosana Hance are found in south- eastern Asia (Bogle 1986). A phylogenetic analysis has recently been conducted based on DNA sequences of the plastid gene matK (Li et al. 1997a). Li et al. found that the western Asian species L. orientalis is most closely related to the New World species L. styraciflua. This result is consistent with an earlier allozyme study (Hoey and Parks 1991). The finding is significant because it suggests a possible phytogeographical connection be- tween the western Eurasian continent and the New World.

The close relationship of Liguidambar styraciflua and L. or- ientalis is moderately well supported by bootstrap values in the matK analysis and no homoplasy was found. However, there are only seven phylogenetically informative characters in the data set. Therefore, it is desirable to gather more evidence to test this phy- logenetic hypothesis. In this note, we report recent progress.

Three more regions of DNA have been sequenced, including the ITS (Internal Transcribed Spacer) region of nuclear ribosomal DNA (Baldwin et al. 1995), the intron of chloroplast gene trnL (Taberlet et al. 1991), and the exon 9-exon 12 region of the GBSS (Granule-Bound Starch Synthase) gene (Dai et al. 1996; Mason and Kellogg, unpubl. data).

DNA extraction, sequencing reactions, and PCR (Polymerase Chain Reaction) amplification for ITS were conducted as de- scribed in Li et al. (1997b, 1997c). PCR amplification of the trnL intron was carried out using primers c and d of Taberlet et al. (1991) with a thermocycler program of 30 cycles of 94°C for 35 sec., 55°C for 30 sec., and 72°C for 95 sec. The final cycle was

87

88 Rhodora [Vol. 101

followed by a seven minute extension at 72°C. PCR amplification of the GBSS gene was conducted using primers GBSSF2 (5’TGGCATGGATACCCAAGAGT3’) and GBSSR2 (5’'CCTTC- TTTCACAGTGTCAAC3’). We have successfully amplified a fragment of about 800 base pairs for several flowering plant taxa, including Hamamelis (Hamamelidaceae), Stewartia (Theaceae), and Viburnum (Adoxaceae). The thermocycler program consisted of a 60 sec. hotstart at 96°C and 40 three-temperature cycles, followed by a 15 minute extension at 72°C. Each cycle had de- nature and extension temperatures of 96°C for 60 sec. and 72°C for 90 sec., respectively. The annealing temperatures were 56°C, 54°C, and 52°C for the first two cycles, the second two, and the remaining 36 cycles, respectively. The annealing time was 60 sec. for all cycles. Sequencing reactions were carried out using Amer- sham cycle sequencing kit (Amersham Life Science Inc., Arling- ton Heights, IL) and following the manufacturer’s instructions. Sequences were determined with an ABI 377 automated sequenc- er (Applied Biosystems, Inc. Foster City, CA). We obtained 348 base pairs from the ITS region (partial 5.8S plus ITS-2), 524 sites from the frnL intron, and 776 base pairs from the GBSS gene.

The parsimony analyses were conducted using the exhaustive tree search option of PAUP 3.1.1 (Swofford 1993). Trees were rooted using the same outgroup, Mytilaria laosensis Lecompte, as in the previous study (Li et al. 1997a), except that Exbucklan- dia R. W. Br., which is closely related to Mytilaria (Li et al. unpubl.), was used for the GBSS data set because we were unable to amplify the GBSS gene for Mytilaria due to its genomic DNA deterioration. Decay analysis (Donoghue et al. 1992) and 1000 bootstrap replicates (Felsenstein 1985) were carried out to indi- cate the relative support for the clades. Characters were unordered and unweighted, and gaps were treated as missing data.

The ITS data contained 82 variable sites, 15 of which were phylogenetically informative. The interspecific divergences be- tween Liquidambar species ranged from 0.9-6.6%. A single most parsimonious tree was generated based on the ITS data set with a consistency index of 0.99. The tree was comprised of two clades, one of which contained L. acalycina and L. formosana and was supported by a bootstrap percentage of 100% and a de- cay index of nine steps. The other clade was composed of L. orientalis and L. styraciflua and was not strongly supported, with bootstrap and decay values of 69% and one step, respectively.

1999] Note 89

EH L. styraciflua

«JTS x -- matK 0 -- trnL intron --GBSS___OGS -- outgroups

Figure 1. The single most parsimonious tree of 535 steps of Liquidambar based on sequences of ITS, matK, GBSS, and trnL intron. CI = 0.98. Num-

bers above and below the branches are bootstrap percentages and decay index values, respectively. Symbols represent unambiguous, potentially informative changes of each data set along the branches.

The trnL intron data set had 21 variable sites, two of which were phylogenetically informative. The sequence divergences be- tween Liquidambar species were from 0—0.8%. Liquidambar acalycina and L. formosana had identical trnL intron sequences. The parsimony analysis generated two equally short trees, one of which showed the tree topology produced by the ITS data, while the other tree did not resolve the relationships of L. formosana, L. acalycina, and the clade of L. orientalis and L. styraciflua. The consistency index was 1.0.

There were 105 variable sites in the GBSS data set, 15 of which were informative. Parsimony analysis resulted in one single short- est tree of 111 steps, with a consistency index of 0.96. In the

90 Rhodora [Vol. 101

phylogenetic tree, eight and seven informative sites supported the clade of Liguidambar orientalis—L. styraciflua and L. acalycina— L. formosana, respectively. Bootstrap values for the two clades were 78% and 82%, respectively.

The four data sets were congruent, including matK, ITS, trnL intron, and GBSS, and the combination of them created a data set

164 characters. The parsimony analysis, using both Zxbuck- landia and Mytilaria as outgroups with Mytilaria GBSS sequenc- es coded as missing data, produced a single most parsimonious tree with a consistency index of 0.98. The gE tree showed the same species relationships as described in al (1997a). Both bootstrap percentages and decay values were high, 100% and 18 steps for the L. acalycina—L. formosana clade, and 86% and three steps for the clade of L. orientalis and L. styra- ciflua. Figure 1 shows the number of unambiguous changes from each of the four data sets that support the two clades.

This follow-up study strongly substantiates the previous hy- pothesis that the western Asian species Liguidambar orientalis is more closely related to the New World species L. styraciflua than to the southeast Asian species. Additionally, we conclude that sequences of the GBSS gene, especially the introns, provide an- other informative nuclear marker (besides nrDNA ITS) in resolv- ing phylogenetic relationships among closely related species.

ACKNOWLEDGMENTS. We thank Robie Mason-Gamer and Toby Kellogg for assistance in designing GBSS gene primers. This study was partially supported by a Putnam Postdoctoral Fellow- ship through the Arnold Arboretum of Harvard University to JL.

LITERATURE CITED

BaALDwiINn, B. G., M. J. SANDERSON, J. M. Porter, M. FE WOIJCIECHOWSKI, AND - J. DONOGHUE. 1995. The ITS region of nuclear ribosomal DNA: A valuable apt > evidence on angiosperm phylogeny. Ann. Missouri Bot. Gard. 82: ce BocLe, A. L. 1986. pee floral morphology and vascular anatomy of the Ha- mamelidaceae: Subfamily Liquidambaroideae. Ann. Missouri Bot. Gard. 73: 325-347.

Dat, W., W. DENG, W. Cul, S. ZHAO, AND X. WANG. 1996. Molecular cloning and sequence of potato granule-bound starch synthase gene. Acta Bot. Sin. 38: 777-784.

DonoGuHuE, M. J., R. G. OLMsTEAD, J. E Smitu, and J. D. PALMER. 1992.

1999] Note 91

Phylogenetic relationships of Dipsacales based on rbcL sequences. Ann. Missouri Bot. Gard. 79: 333-345

FELSENSTEIN, J. 1985. Confidence limits on phylogeny: An approach using

the bootstrap. Evolution hy 783-791.

Hoey, M. T. AND C. R. Parks. 1991. Isozyme divergence between eastern Asian, North American, Cid Turkish species of Liguidambar (Hama- ae Amer. J. Pos 78: 938-947.

Li, J., A. L. BOGLE, AND A. S. KLEIN. 1997a. Interspecific SRT and genetic divergence of the disjunct genus Liqguidambar (Hamamelidaceae) inferred from DNA sequences of plastid gene matK. Shindinn’ 99: 229-

0

——-, AND 997b. Phylogenetic relationships in the Cor- ylopss —— (Hamamelidaceae): Evidence from sequences of the in- bed spacers of nuclear ribosomal DNA and morphology. Rhodora 99. 302 318

AND K. PAN. 1997c. Close relationship between

Shandpdeniivon and Parone (Hamamelidaceae), evidence from ITS se-

quences of nuclear ribosomal DNA. Acta Phytotax. Sin. 35: 481—493. SworFForD, D. L. 1993. PAUP: Phylogenetic Analysis Using Parsimony, ver-

sion 3.1.1. AOE ts of Molecular Systematics, Smithsonian Institu-

tion, Washington, Peer, P, L. GrELty, fi PauTOu, AND J. Bouvet. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. PI.

Molec. Biol. 17: 1105—1109.

RHODORA, Vol. 101, No. 905, pp. 92-94, 1999 BOOK REVIEW

The Savage Garden: Cultivating Carnivorous Plants by Peter

"Amato. 1998. xxii + 314 pp. more than 200 photographs

and illustrations, most in color. ISBN 0-89815-915-6 $19.95 (paperback). Ten Speed Press, Berkeley, CA.

Carnivorous plants, a group of about 500 species and numerous cultivars, have captivated botanists, horticulturalists, and students of all ages since Darwin (1875) wrote one of the earliest books on the group. The more easily grown species and those locally abundant in boggy areas have been widely employed by teachers to stimulate interest in young students in plant biology and ecol- ogy. Unfortunately, until recently, little information has been available in book form in public and college libraries for teachers and students who want to know more about growing these plants. In recent years, three books with a worldwide scope have been published and widely distributed: Pietropaolo and Pietropaolo (1986), Lecoufle (1989, in translation), and Cheers (1992). These three works cover some of the same ground as the earlier and excellent books by Slack (1980, 1988) which are difficult to ob- tain.

Peter D’Amato’s book surpasses these earlier works in the sheer volume of cultivation information and lore provided for individual species and cultivars. It is based on Peter’s many years of experience with growing these plants; of selling them through his greenhouse, mail order and Internet nursery, California Car- nivores; and of making numerous presentations at schools, hob- byist meetings, and on television. His goal, as he described it during a talk at the first meeting of the International Carnivorous Plant Society in Atlanta in 1997, was to popularize the plants and show how they can be grown in a wide variety of ways indoors and out. The range of creative, and sometimes whimsical, ways to grow these plants, well illustrated and described in the book, is its strongest feature. Since the book covers species worldwide, it also provides a starting point for exploring plant diversity and plant geography. The amazing radiation of Drosera species in Australia, and the narrowly restricted endemic genera Darling- tonia, Heliamphora, and Cephalotus which use trapping mecha- nisms similar to Sarracenia and Nepenthes are good examples.

92

1999] Book Review 93

A few minor faults should be noted. Many of the color pho- tographs are small, but this allows for more text to flow around the figures creating a tighter integration of the two. On the other hand, this design foregoes including any large in situ shots such as those in Schnell’s (1976) Carnivorous Plants of the United States and Canada or Clarke’s (1997) truly stunning book on the Nepenthes of Borneo. As noted in the review published in the September (1998) issue of the Carnivorous Plant Newsletter, there are some spelling errors and some of the cultivars listed are not well documented in the literature. Readers interested in an overview of the unique physiological and ultrastructural features of carnivorous plants may wish to supplement this book with the earlier work by Juniper et al. (1989). Some of the guidelines presented for growing individual species may be less effective outside the northern California climate where the author lives. Apparently, no hardcover edition is available.

Overall, the well named Savage Garden is a bargain for the wealth of information it contains. As a CP enthusiast, I use it often. It may well achieve “‘bible’’ status for growers of these plants, especially as tissue culture (described in the first Appendix by expert Rob Gagliardo) makes more species available to the general public. I would recommend this book for all libraries, and to botanists and horticulturalists as a one-book gateway to know- ing more about these fascinating plants.

LITERATURE CITED

Cueers, G. 1992. A Guide to the ‘nail Plants of the World. HarperCollins Publishers, New Yor

CLARKE, C. 1997. Nepenthes of a Natural History Publications, Kota Kinabulu, Sabah, Malaysia

Darwin, C. 1875. fanectiocninns Plants. John Murray Publishers, London.

JuNIPER, B. E., R. J. RoBins, AND D. M. JoEL. 1989. The Carnivorous Plants. Academic Press, London.

LECOUFLE, M. 1989. Carnivorous Plants: Care and Cultivation. Cassell Vil- liers House, London. (English translation.)

MEyYERS-RICE, B. 1998. Book review. Carniv. Pl. Newslett. 27: 72-7

rete J. AND P. eas 1986. ease etsivit Plants of the World.

mber Press, Portland, O

Papers D. E. 1976. hae Plants of the United States and Canada.

John E Blair Publisher, Winston-Salem, NC.

94 Rhodora [Vol. 101

SLack, A. 1980. Carnivorous Plants. MIT Press, Cambridge, MA. —. 1988. Insect-Eating Plants and How to Grow Them. University of Washington Press, Seattle, WA.

—DAvID LANE, Biological Sciences Library, Kendall Hall, Uni- versity of New Hampshire, 129 Main St., Durham, NH 03824- 3590.

RHODORA, Vol. 101, No. 905, pp. 95-100, 1999 NEBC MEETING NEWS

October 1998. Dr. Lisa A. Standley, the Club’s Vice President, spoke on the topic ‘‘Beyond the Brooks Range—Flora and Fauna of the Arctic National Wildlife Refuge.’”” Among Dr. Standley’s many activities, we learned, is participating in Sierra Club out- ings. Twice in recent years, she has enjoyed ten-day backpacking trips to the Arctic National Wildlife Refuge in northeastern Alas- ka. Both trips were in mid-June, and started with a flight into Fairbanks and transfer to a smaller plane that flew through passes in the Brooks Range to the Romanzof Mountains and the coastal plain of the Beaufort Sea, a couple of hundred miles north of the Arctic Circle. The area hiked was between two rivers, the Jago and the Aichilik, which flow northward from the mountains, crossing the coastal plain to the sea. The Refuge is contiguous with Indian lands and National Parks, which add to the wilderness landscape. It is home to the caribou’s Porcupine Herd calving grounds and their migration routes to the mountains. Fortunately for us, Lisa was armed with a good camera and the ability to use it well. We were treated to excellent images of the region’s plant life, interspersed with those of the often present and possibly curious caribou. Also, landscape shots illustrated some of the Ref- uge’s varied habitats and unadulterated beauty. The hiking area ranged in elevation from near sea-level to around 5000 ft. Most of the landscape is devoid of tall trees, although some of the narrower valleys supported white and black spruce in sheltered areas. Low willows were the more typical woody vegetation. The narrow valleys generally run east-west while larger river valleys run north-south. Precipitation is surprisingly low there, with only about 10 inches per year, Standley said. The few glaciers seen while crossing the Brooks Range were relatively small and not growing, evidently remnants of earlier times with higher rates of precipitation.

The slide images gave a good sampling of the dominant plant families in the Refuge and the Arctic region, in general. The Cyperaceae, a family Standley knows especially well because of her research on the genus Carex, is one of them. Sedges were well represented and tipsy tussocks of cottongrass were frequently underfoot. More frustrating for Lisa than the tipsy tussocks, per- haps, was that nearly all the sedges present in June were flowering rather than fruiting, making identification very challenging. An-

95

96 Rhodora [Vol. 101

other family well represented was the Salicaceae. A favorite for Standley was Salix minima which stood less than an inch tall with catkins of reddish flowers. The Saxifrage family was rep- resented by several species of Saxifraga including S. oppositifol- ia, a circumboreal species present in New England, and the very unusual S. eschscholtzii with its tiny cushion-like rosettes of suc- culent leaves only 2 mm across. Jumping to the Rose family, we saw Potentilla hypartica (or P. nana, in some books), a close relative of New England’s federally endangered P. robbinsiana. Also representing the Rosaceae were both species of Dryas. The hikers liked seeing Dryas, since it meant they would be walking on gravel substrate and not wobbly tussocks. Ericads were also represented in the tundra by white flowering Cassiope [or Har- rimanella in some references], pink flowered Rhododendron lap- ponicum described by Standley as weedy everywhere, and Loise- leuria seen at around 5000 ft. elevation.

Also illustrated by Standley were: a Douglasia species (Pri- mulaceae) which is endemic to Alaska and the Yukon; a Hedy- sarum (Fabaceae) which has aromatic, edible roots eaten by griz- zly bears; yellow poppies, which trap heat and attract flies in cup- like flowers that tilt toward the sun, which in June’s solstice sky shines for 24 hours per day; nitrophilous, orange-colored lichens growing on rocks where birds perch; caribou trails made in the tundra in the 1940s; caribou skeletons used for drying wet socks; musk-oxen simulating “fringed sofas” swaying in the breeze; ae- rial views of river meanders revealing a hundred or more years of geomorphology; vertical Jurassic formations with marine fos- sils; cliffs with gyrfalcon nest sites: sloping bogs at 4000-5000 ft.; and a grizzly sow with cub.

Standley recommended Pielou’s Arctic Naturalist and Birds of

laska.

1999] NEBC Meeting News 97

nine sample sites within a continuous, unfragmented habitat of the Presidential Range in the White Mountains of New Hamp- shire. Previous studies on the effects of fragmentation on loss of genetic diversity, he said, have involved habitats with a relatively recent history of fragmentation (i.e., less than a few hundred years) and only one taxon. He thought, by examining high peak populations presumably separated for thousands of years, that the effects of time on genetic drift and genetic diversity might be more apparent. By examining three species, he hoped to reduce the possibility of any erroneous conclusions made by assuming that what is true for one is true for all. Lindwall identified the three key questions he wanted to answer in the study as: 1) Do fragmented plant populations in the Adirondack peaks have less diversity than the continuous population in the White Mountains? 2) Is there more gene flow in the Presidential Range than among the fragmented populations in the Adirondacks? 3) What effect does greater habitat area have on diversity in the White Moun- tains versus the smaller area for each of the isolated Adirondack sites?

A fortuitous coincidence of Lindwall’s site design, he added, was that the overall land area and distances between sites for the two study areas were approximately the same. To quantify the relative abundance and frequency of each species, 6000 plots, each one m2, were examined. The genetic diversity was assessed using allozyme analysis. The three species studied were Minuar- tia groenlandica, which appears to be exploiting disturbed trail- edge habitat, Carex bigelowii, which forms large patches in the White Mountains, and Diapensia lapponica, a monotypic genus found in tundra. Three thousand tissue samples were taken and analyzed during the study.

For each ines Lindwall’s three questions, the answers were ‘“‘ves,”” “no,” and “‘maybe.’’ Did the fragmented Adirondacks have lower genetic diversity? For Diapensia lapponica, the an- swer was a Statistically significant “tyes,” but for Minuartia groenlandica, he found higher diversity at all Adirondack peak sites than at the Presidential Range subsites. The results for Carex bigelowii were not as easy to interpret. The overall genetic var- iability was higher in the White Mountains, but because C. bi- gelowii is less abundant in the Adirondacks than the Presidential Range, the sample size was small and only one of four indices was higher, statistically. Thus, we have a “‘maybe.”” What about

98 Rhodora [Vol. 101

gene flow? Lindwall created dendrograms to illustrate degrees of similarity (or difference) in both genetics and geographic dis- tances among the populations. Comparing Nei’s index of genetic identity for each of the three species relative to the Adirondacks and Presidentials, the answers were again mixed. For C. bigelo- wii, there was a close relationship among all sites in New Hamp- shire but not so among the New York sites. Minuartia groenlan- dica, on the other hand, showed no particular pattern with gen- erally good gene flow across the board. However, the most ge- netically distant population in the Adirondacks was from the most distantly isolated peak, the Gothics. The story with D. lapponica also seemed to relate to distance between sites. In both areas there appeared to be good gene flow with near neighbors, such as among the four McIntyre Ridge peaks in the Adirondacks, but less so when distance was greater between sites. What role does habitat area play? With C. bigelowii, there was a clear relation- ship: bigger places had more variability. Just the Opposite was true for Minuartia: the smaller sites in the Adirondacks had sta- tistically higher variability than the continuous population in the Presidentials. For Diapensia, size appeared to have no effect, and thus we have a “‘maybe”’ answer.

There was one general conclusion that fit all three species, Lindwall said in summary: The greatest amount of genetic di- versity occurs where each species is the most abundant. He also concluded that we should neither assume that species will behave the same despite similar histories, nor for conservation planning purposes assume that the largest habitat area will support the most diverse population of a given species.

December 1998. The program, entitled “Verdant Venues and Ventures: Visible and Verbal Visions” represented the annual event where Club members are invited to make short presenta- tions on their explorations over the year. Keith Williams led off with images from a South America vacation trip with his wife in May. It was the middle of the dry season in Brazil, their first destination, but they still saw lots of water because much of their time was spent on the coast and in the Pantanal, a huge wetland that extends into two other South American countries. Plants fea- tured in the slide images were Tabebuia alba, an endangered tree species in Brazil; Cuphea melvilla, a prolific shrub along the Pan- tanal waterways; and Ludwigia inclinata and Cabomba furcata

1999] NEBC Meeting News 99

growing in sloughs and shallows of the Pantanal. He ended with shots from Peru’s Inca Trail to Machu Picchu and an image of an Equisetum growing from the mortar of stone ruins.

Marsha Salett followed Keith with a brief introduction to her Master’s degree project at the University of Massachusetts—Bos- ton which is to create a CD-ROM version of a natural history guide to bogs of southern New England. She showed images of several bogs with public access that she might feature in the guide, as well as a few that lack easy access or boardwalks that she may omit. Her intent is to present explanations and illustra- tions of bog types and common species such as Kalmia angus- tifolia and Ledum groenlandicum. Dichotomous keys and images of plants in flower and fruit will be provided to help with iden- tifications.

Lois Somers then took us back to the tropics with images of a trip with husband Paul to Costa Rica. Being a registered nurse, not a botanist, she used a few wildlife images to illustrate some of the critters botanists need to be on the watch for while probing the greenery. The images included an orange-kneed tarantula seen in the Monteverde cloud forest and an eyelash viper seen at Brau- lio Carrillo National Park. Aquatic critters to be aware of included caiman seen on the Cano Negro River near the Nicaraguan border and the much larger and fiercer crocodiles of the Palo Verde re- gion.

Joanne Sharpe’s slides started in Costa Rica with an image of Danaea wendlandii, one of the fern species she studied there for six years. She then took us to Puerto Rico for a look at distur- bance studies of ferns in a palm forest before and after Hurricane Georges, and in mangrove swamps, where the 14 ft. tall leather fern, Acrostichum danaeifolium, was regenerating following four years of hydrologic disturbance from dike construction. Her last stop was Maine with images from the Coastal Maine Botanical Gardens in Boothbay, where Nyssa sylvatica can be found at or near its northern limit.

David Hunt continued the regional theme with images from New York where he has been helping to refine the state’s plant community classification, particularly in the Northern Appala- chian Ecoregion. His images included riverside ice meadows with Prunus pumila and Andropogon gerardii and pine-dominated rocky summit communities with either pitch or red pine and as- sociates such as Vaccinium myrtilloides, Amelanchier bartrami-

100 Rhodora [Vol. 101

ana, and Oryzopsis pungens. He then took us underwater at Lake George where he has been doing underwater vegetation sampling at depths up to 40 ft. In shallow bays he found Potamogeton amplifolius—Vallisneria americana and Eriocaulon aquaticum— Elatine americana to be common community types, whereas sandy deltas had associations of Lobelia dortmanna and Myrio- phyllum pinnatum. In deeper waters he found associations of Na- jas flexilis, Potamogeton gramineus, and P. perfoliatus. At 30 ft., he found beds of Jsoetes macrospora and Potamogeton robbinsii, and at 40 ft., a dense cover of Nitella flexilis. With this success, he’s now tackling marine eelgrass environments of Long Island.

The next three presenters came as a team representing the new-

formed Botanical Club of Cape Cod and the Islands. Don Schall spoke about the group’s search for and likely rediscovery of an extant population of Asclepias purpurascens on the Cape and the discovery of water hyacinth, Eichhornia crassipes, thriv- ing in a spring upwelling near a Barnstable cranberry bog. Mario DiGregorio discussed, with a vial sample in hand, the group’s discovery of a county record for Wolffia papulifera from a fresh- water pond in Barnstable and showed images of sandplain grass- land rarities: Liatris scariosa var. novae-angliae being visited by a monarch butterfly, Aster concolor at the northern limit of its range, Aristida purpurascens, New England’s only perennial awn- grass, and Prenanthes serpentaria from Nantucket. Pamela Pol- loni continued with the discussion of P. serpentaria by pointing out its hairy calyx, which distinguishes it from P. trifoliata, and other aspects of its life history such as pollination by Bombus bees and how to recognize the juvenile plants.

—PauL Somers, Recording Secretary.

INFORMATION FOR CONTRIBUTORS TO RHODORA

Submission of a manuscript implies it is not being considered for publication simultaneously elsewhere, either in whole or in part.

GENERAL: Manuscripts should be submitted in triplicate. The text must be double-spaced throughout, including tables, figure legends, and literature citations. Use a non-proportional font throughout and do not justify the right margin. Do not indicate the style of type through the use of capitals, underscoring, or bold, except for names of genera and species which should be in italics or underscored throughout. Do not underline punctuation. All pages should be num- bered in the upper right-hand corner. For guidance in matters not addressed here, consult the editorial office by phone at (603) 862- 3205, FAX (603) 862-4757, or e-mail: janets@christa.unh.edu. Brev- ity is urged for all submissions. Submit manuscripts to the Editor-in- Chief.

TITLE, AUTHOR(S), AND ADDRESS(ES): Center title, in capital letters. Omit authors of scientific names. Below title, include au- thor(s) name(s), affiliation(s), and address(es). If “‘current address” is different, it should follow immediately below, not as a footnote.

ABSTRACT: An abstract and a list of key words should be included with each paper, except for shorter papers submitted as Notes. An abstract must be one paragraph, and should not include literature citations or taxonomic authorities. Please be concise, while including information about the paper’s intent, materials and methods, results, and significance of findings.

TEXT: Main headings are all capital letters and centered on one line. Examples are: MATERIALS AND METHODS, RESULTS, and DIS- CUSSION. Do not title the Introduction. Do not combine sections of the paper (such as Results and Discussion), or use Conclusions or Summary. Second level headings should be indented, bold, upper and lower case, and end with a period. Taxonomic authorities should be cited for all species names at their first usage in the text, or in a referenced table. Cite each figure and table in the text in numerical order. Each reference cited in the text must be in the Literature Cited. Cross-check spelling of author(s) name(s) and dates of publication. Literature citations in the text should be as follows: Hill (1982) or (Hill 1982). For two or more authors, cite as follows: Angelo and Boufford (1996) or (Angelo and Boufford 1996). Cite several refer- ences alphabetically by first author, rather than chronologically. With- in parentheses, use a semicolon to separate different types of citations (Hill 1982; Angelo and Boufford 1996) or (Figure 4; Table 2).

FLORAS AND TAXONOMIC TREATMENTS: Specimen citation should be selected critically, especially for common species of broad

101

102 INFORMATION FOR CONTRIBUTORS

distribution. Keys and synonymy for systematic revisions should be prepared in the style of ‘““A Monograph of the Genus Malvastrum,” S. R. Hill, RHODORA 84: 159-264, 1982. Designation of a new taxon should carry a Latin diagnosis (rather than a full Latin descrip- tion), which sets forth succinctly how the new taxon differs from its congeners.

LITERATURE CITED: All bibliographic entries must be cited in the paper, unless a special exception has been made by the Editor (such papers will be allowed a REFERENCES section). Verify all entries against original sources, paying special attention to spelling and de- tails of publication. Cite references in strict alphabetical order by first author’s surname. Do not write authors’ names in all capital letters. References by a single author precede multi-authored works of same senior author, regardless of date. Use a long dash when the author(s) is the same as in the entry immediately preceding (see recent issues). Refer to Botanico-Periodicum-Huntianum (B-P-H 1968) and B-P-H/ Supplement (1991) for standardized abbreviations for journals.

TABLES: Tables must be double-spaced. Tables may be continued on an extra page, if necessary. As much as possible, the title should be self-explanatory. Do not use footnotes; instead, add notes after the end of the table title. Broadside tables should be avoided, if pos- sible. Each table should be cited in the text in numerical order.

URES: Illustrations must be either black and white half-tones (photographs), drawings, or graphs. Illustrations must be camera- ready; flaws cannot be corrected by the Editor or the printer. Add symbols or shading with press-on sheets. The printed plate will be 4 X 6 inches; be sure that illustrations are proportioned to reduce correctly. Allow space for a caption, if possible. Magnification/re- duction values should be calculated to reflect the actual printed size. Maps must indicate scale and compass direction. The double-spaced list of legends for figures should be provided on a separate page. Each figure should be cited in the text in numerical order.

THE NEW ENGLAND BOTANICAL CLUB 22 Divinity Avenue Cambridge, MA 02138

The New England Botanical Club is a nonprofit organization that promotes the study of plants of North America, especially the flora of New England and adjacent areas. The Club holds regular meetings, and has a large herbarium of New England plants and a library. It publishes a quarterly journal, RHO- DORA, which is now in its 101st year and contains about 400 pages per volume. Visit our web site at http://www.herbaria. harvard.edu/nebc/

Membership is open to all persons interested in systematics and field botany. Annual dues are $35.00, including a subscrip- tion to RHODORA. Members living within about 200 miles of Boston receive notices of the Club meetings.

To join, please fill out this membership application and send with enclosed dues to the above address.

Regular Member $35.00 Family Rate $45.00 Student Member $25.00

For this calendar year For the next calendar year ee

Name

Address

City & State Zip

Phone FAX

email

Special interests (optional):

Ann a ing

3 1 753 00364 9

THE NEW ENGLAND BOTANICAL CLUB

Elected Officers and Council Members for 1998-1999:

President: David S. Conant, Department of Natural Sciences, yndon State College, Lyndonville, VT 05851

Vice-President (and Program Chair): Lisa A. Standley, Vanasse Hangen Brustlin, Inc., 101 Walnut St., PO. Box 9151, Wa- tertown, MA 02272

Corresponding Secretary: Nancy M. Eyster-Smith, Department of Natural Sciences, Bentley College, Waltham, MA 02154- 4705

Treasurer: Harold G. Brotzman, Box 9092, Department of Bi- ology, Massachusetts College of Liberal Arts, North Adams, MA 01247-4100 Recording Secretary: Paul Somers Curator of Vascular Plants: Raymond Angelo Assistant Curator of Vascular Plants: Pamela B. Weatherbee Curator of Nonvascular Plants: Anna M. Reid Librarian: Leslie J. Mehrhoff Councillors: W. Donald Hudson, Jr. (Past President) Michael J. Donoghue 1999 Arthur V. Gilman 2000 Karen B. Searcy 2001 Matthew Hickler (Graduate Student Member) 1999

Appointed Councillors: David E. Boufford, Associate Curator Janet R. Sullivan, Editor-in-Chief, Rhodora

Journal of the New England Botanical Club

CONTENTS

Inventory and vegetation classification of floodplain forest communities in M

chusetts. Jennifer B. Kearsley 105 Jaltomata lojae (Solanaceae): Description and floral — of a new An- ean species. Thomas Mione and Luis A. Seraz 136 The reproductive biology of Magnolia evn Larry K. Allain, Mi- Zavada, and Douglas G. Matth 143 The taxonomy of Carex section Scirpinae (Cyperaceae). Debra A. Dunlop and Garrett E. Crow 163 NEW ENGLAND NOTE Rare and non-native plants of Massachusetts’ floodplain forests. Jennifer Kearsley 200 BOOK REVIEW Wild Orchids Across North America: A Botanical Travelogue .......... 206 NEBC MEETING NEWS 208 Information for Contributors 213 NEBC Membership Form 215 NEBC Officers and Council Members inside back cover MISSOURI BOTANICAL JUL 1 3 1999 GARDEN LIBRARY Vol. 101 Spring, 1999 No. 906

Issued: June 29, 1999

The New England Botanical Club, Inc.

22 Divinity Avenue, Cambridge, Massachusetts 02138

RHODORA

JANET R. SULLIVAN, Editor-in-Chief

Department of Plant Biology, University of New Hampshire, Durham, NH 03824

ANTOINETTE P. HARTGERINK, Managing Editor

Department of Plant Biology, University of New Hampshire, Durham, NH 03824

Associate Editors

HAROLD G. BROTZMAN STEVEN R. HILL DAVID S. CONANT THOMAS D. LEE GARRETT E. CROW THOMAS MIONE

K. N. GANDHI—Latin diagnoses and nomenclature

RHODORA (ISSN 0035-4902). Published four times a year (January, April, July, and October) by The New England Botanical Club, 810 East 10th St., Lawrence, KS 66044 and printed by Allen Press, Inc., 1041 New Hampshire St., Lawrence, KS 66044-0368. Periodicals postage paid at Lawrence, KS. POSTMASTER: Send address changes to RHODORA, PO. Box 1897, Lawrence, KS 66044-8897.

RHODORA is a journal of botany devoted primarily to the flora of North America. Monographs or scientific papers concerned with systemat- ics, floristics, ecology, paleobotany, or conservation biology of the flora of North America or floristically related areas will be considered.

ACCREDITED with the International Association for Plant Taxonomy for the purpose of registration of new names of vascular plants (ex- cluding fossils).

SUBSCRIPTIONS: $75 per calendar year, net, postpaid, in funds paya- ble at par in United States currency. Remittances payable to RHO- DORA. Send to RHODORA, PO. Box 1897, Lawrence, KS 66044- 8897.

MEMBERSHIPS: Regular $35; Family $45; Student $25. Application

form printed herein

BACK ISSUES: Questions on availability of back issues should be ad- dressed to Dr. Cathy A. Paris, Department of Botany, University of Vermont, Burlington, VT 05405-0086. E-mail: cparis @ moose. uvm.edu.

ADDRESS CHANGES: In order to receive the next number of RHO- DORA, changes must be received by the business office prior to the first day of January, April, July, or October.

RHODORA, Vol. 101, No. 906, pp. 105-135, 1999

INVENTORY AND VEGETATION CLASSIFICATION OF FLOODPLAIN FOREST COMMUNITIES IN Missoury BOTANICAL MASSACHUSETTS

JENNIFER B. KEARSLEY JUL 0 . 1999

Massachusetts Natural Heritage & Endangered Species Program, Massachusetts Division of Fisheries & aa Route 13.BARDEN LIBRARY Westborough, MA 015

ABSTRACT. Floodplain forests on eleven rivers in Massachusetts were sur- veyed to determine the variation in vegetation and soils across a range of hydrologic, physiographic, and climatic conditions. Quantitative vegetation data collected from 124 plots at 43 sites were analyzed using SPAN and DECORANA (DCA), and six community types were identified. The six

pes were: Type I—Riverine island floodplain forests (Acer saccharinum— Populus deltoides-Acer negundo—Matteuccia struthiopteris association);

Type II—Major-river floodplain forests (A. saccharinum—P. deltoides—Lapor- tea canadensis association); Type I1I—Transitional floodplain forests (A. sac- charinum—Arisaema dracontium association); Type [V—Small-river flood- plain forests (A. saccharinum—Fraxinus pennsylvanica—Quercus palustris as- sociation); Type V—Alluvial swamp forests (Acer rubrum—A. saccharinum— Q. bicolor association); and 1 VI—Alluvial terrace forests (A. rubrum—

densis, Boehmeria cylindrica, and Onoclea sensibilis. Results of the classi-

rang: showed variation in floodplain forest vegetation composition among ers in Massachusetts corresponding to significant differences in soil mot-

da soil texture, presence/absence of a surface organic layer, and soil pH.

Key Words: Acer saccharinum, community classification, DECORANA, floodplain forest, Massachusetts, ordination, TWINSPAN

Floodplain forests, which develop on alluvial mineral soils within the zone of active flooding of rivers and streams, are con- sidered to be among the most threatened, globally significant wet- land community types in New England. Due to their high soil fertility and scenic qualities, floodplain forests have largely been converted to agriculture or lost to housing and industrial devel- opment. While several studies have addressed the relationship between floodplain forest vegetation and environmental variables within a single site or river basin in New England (Metzler and

amman 1985; Veneman and Tiner 1990), this study addresses the variability in floodplain forest vegetation and environments across river basins and physiographic regions. The objectives of

105

106 Rhodora [Vol. 101

the current study were to conduct a statewide vegetation classi- fication of floodplain forest communities, to determine the distri- bution of defined community types across drainage basins and rivers, and to assess differences in environmental parameters among the identified floodplain forest community types.

The Massachusetts inventory and classification work is part of a regional effort to classify floodplain forests by state Natural Heritage Programs and The Nature Conservancy. Results of these projects will provide the baseline community data necessary for future in-depth studies of floodplain forest communities, and for land protection and conservation of these ecologically significant wetland communities.

MATERIALS AND METHODS

Site selection. Potential floodplain forest sites were identified using USGS topographic quadrangles, Natural Resource Conser- vation Service soil surveys, and color-infrared (CIR) aerial pho- tography. A combination of 1:25,000 scale, leaf-on CIR aerial photography from an unpublished community inventory of the Connecticut River Valley (Motzkin 1993), and 1:12,000 scale, leaf-off CIR aerial photography obtained from the Massachusetts Department of Environmental Protection Wetlands Conservancy

ogram were used. Potential floodplain forest sites were identi- fied using the following criteria: (1) low, forested sections of greater than 3 ha occurring within 1—2 contour intervals (10—20 ft. elevation) of river’s edge; (2) presence of alluvial soils; and (3) evidence of spring flooding and forest vegetation on aerial photography. The Massachusetts Natural Heritage and Endan- gered Species Program Biological and Conservation Database was also used to locate potential floodplain forest sites by iden- tifying localities of tracked, state-protected rare species known to occur in floodplain forest habitats.

Using the resources and criteria listed above, 144 potential floodplain forest sites were identified in the state. Based on pre-

a

Figure 1. Massachusetts’ rivers and sub-ecoregions with sites surveyed for floodplain forest vegetation classification. Sub-ecoregions containing sur- vey sites are shaded in grey.

SELECTED SUB-ECOREGIONS OF MASSACHUSETTS A - Western New England Marble Valleys

B - Connecticut Valle

C - Southern New England Coastal Plains

D - Narragansett & Bristol Lowland

-_— — (5 9 «/

N River A Ecoregion Boundary

/V Sub-ecoregion Boundary

Or

[6661

SONMUNUIUIOD 1S910.J UTe[dpoo],j—Aoz]sieay

108 Rhodora [Vol. 101

Table 1. Drainage basins and river sections with sampled floodplain forest communities. Tributaries refer to third order or smaller streams. 50% Ex- ceedance values indicate the discharge of 50% of flows annually, averaged over the period of record (Socolow et al. 1995). Exceedance values are given in cubic feet per second (cfs).

Drainage 50% Num- Basin Exceed- Num-_ ber Area (sq. ance berof of

Basin River miles) (cfs) Sites Plots Blackstone Blackstone 25.6 422 1 1 Connecticut Connecticut at 9,660 11,000 17 31

Connecticut tributaries — — 7 24

Deerfield 557 950 7 11

Housatonic Housatonic at 465 456 5 13

Ashley Falls, MA

Ipswich Ipswich 44.5 37 1 4 Merrimack Assabet 116 125 1 4 Concord 400 481 1 4

Merrimack 4,635 5,110 4 4

Nashua 435 365 2 43

Nashua tips — — 1 3

Shawshee 36.5 38 2, 6

Taunton SERIE 84.3 113 1 6 Totals 43 124

liminary field checks of potential sites, 55 were found to be semi-natural forested floodplain sites with evidence of periodic flooding (e.g. floodlines on trees, flood debris, or scoured sur- faces) and a relative lack of evidence of human disturbance (e.g. limited clearings or non-native plant species). Quantitative veg- etation and environmental data were collected at 43 of the semi- natural forested floodplain sites that were distributed across eleven rivers and four physiographic provinces, or sub-ecore- gions (Figure 1).

e eleven rivers ranged in drainage basin area from 25— 10,000 square miles, and in mean 50% exceedance values from 30—11,000 cubic feet per second (cfs; Table 1). Fifty percent ex- ceedance values are used as indicators of average river discharge; they indicate the minimum discharge in cfs that 50% of all flows exceed annually, averaged over the period of record (Socolow et al. 1995). Identified floodplain forest sites ranged in size from 1

1999] Kearsley—Floodplain Forest Communities 109

to 30 ha. Five sites less than the minimum size criterion of 3 ha were included because they either occurred on state-owned land with easy access (3 sites) or occurred on the Merrimack River 2 sites) where potential sampling sites were limited.

Study area. The eleven rivers sampled in this study are lo- cated within four subregions of the two ecological regions, or ecoregions, occurring in Massachusetts: the Northern Highlands Ecoregion and the Northeastern Coastal Zone (Figure 1; Griffith et al. 1994). These ecoregions are defined as areas with distinct geology, landforms, soils, vegetation, climate, wildlife, water, and human influences (Griffith et al. 1994).

The Northern Highlands Ecoregion includes all of Massachu- setts west of the Connecticut River Valley and the Worcester Pla- teau in north-central Massachusetts as well as most of northern New England and the Adirondack Mountains in New York (Grif- fith et al. 1994). It roughly corresponds to the Adirondack-New England mixed forest-coniferous forest—alpine meadow province described by Bailey (1995). The Northeastern Coastal Zone in- cludes eastern and coastal Massachusetts, most of southern New England, and coastal regions of New Hampshire and southern Maine (Griffith et al. 1994). It falls within the central Appalachian broadleaf forest—-coniferous forest-meadow province described by Bailey (1995).

The lower Housatonic River runs through the Western New England Marble Valleys subregion of the Northern Highlands Ecoregion (Figure 1). Bedrock in this region, also known as the Berkshire Valley, consists of calcitic and dolomitic marbles and limestones; surface water alkalinity values in the area are high

1000 yeq/L; Griffith et al. 1994). The Connecticut and Deer- field Rivers and the lower reaches of their tributaries are included in the Connecticut Valley subregion of the Northeastern Coastal Zone (Figure 1). The Connecticut Valley is characterized by thick outwash, alluvial, and lake bottom deposits overlaying sedimen- tary bedrock. Surface water alkalinity values are generally above 500 wed/L.

The Blackstone, Concord, Assabet, Merrimack, Shawsheen, Ipswich, and Nashua Rivers occur within the Southern New Eng- land Coastal Plains and Hills subregion (Figure 1). This is the largest subregion in southern New England and is variable in its topography and bedrock. Bedrock types in the subregion are

110 Rhodora [Vol. 101

mostly granites, schist, and gneiss, and surface water alkalinity values are generally lower than in the Connecticut Valley, ranging from less than 50 to 500 peq/L. The Threemile River occurs in the Narragansett Bristol Lowland subregion (Figure 1). The Nar- ragansett Basin is similar to the Coastal Plains and Hills subre- gion, but bedrock outcrops are not common, and thick glacial till and outwash deposits cover the area. Surface water alkalinity val- ues are generally between 100 to 300 wedq/L, but several areas have values less than 50 peq/L (Griffith et al. 1994).

Field methods. Vegetation was sampled in 10 m X 20 m (0.02 ha) rectangular plots placed along transects perpendicular to the river. At most sites, two or more transects were placed at least 50 m apart. Each transect was walked and changes in to-

aphy and vegetation were described. A plot was placed with- in each identified topographic or vegetation unit. In small flood- plain forests (=3 ha), one or two plots were subjectively placed within the “‘typical’’ vegetation type(s) and not along transects. The number of plots per site ranged from 1 at small sites to 8 at large sites.

Plots were placed with their long axis parallel to the river. Percent cover of trees (stems >10 cm DBH), shrubs (stems <10 cm DBH), saplings, and vines was visually estimated within each 0.02 ha plot, and percent cover of herbs and seedlings was vi- sually estimated within two 0.0004 ha (2 m X 2 m) square sub- plots. Herbaceous taxa (<1 m tall) occurring within the 0.02 ha plot, but not within the subplots, were also recorded. Nomencla- ture follows Kartesz (1994). Percent cover for all taxa was esti- mated using a modified Braun-Blanquet cover scale with the fol- lowing cover classes: r (single occurrence), <1%, 1-5%, 6-10%, 11-20%, 21-25%, 26-35%, 36-45%, 46-50%, 51-55%, 56— 65%, 66-75%, 76-85%, 86-95%, and 96-100%. The average height and average percent cover of each vegetation stratum were also visually estimated and recorded.

Vegetation data from 124 plots were included in the vegetation classification (Table 1). Eighty-nine plots were surveyed between July and September, 1997, using the methods described above. Existing data from 35 plots collected with equivalent methodol- ogies by other sources were included in the vegetation classifi- cation. Those data were: 10 plots from the Deerfield River (Thompson and Jenkins 1992), 16 plots from the Nashua River

1999] Kearsley—Floodplain Forest Communities 111

(Searcy et al. 1993), 7 plots from the Connecticut River and its tributaries (Motzkin 1993, 1995; Massachusetts Audubon Society, unpubl. data), and 2 plots from the Ipswich River (Massachusetts Audubon Society, unpubl. data).

Environmental data were collected from the 89 plots sampled in 1997. At each plot, one 60 cm deep soil pit was dug and the following soil characteristics were described: depth, soil texture, and color of horizons; depth to mottling; color of mottles; depth of root penetration; and average pH of the mineral soil. The fol- lowing environmental data were also collected for each plot: to- pographic position (terrace, levee, level floodplain, depression), height of floodlines, and the number of stumps, and uprooted and snapped trees. Any evidence of disturbance or land use was also noted.

Data analysis. Vegetation cover data were analyzed using two-way indicator species analysis (TWINSPAN) and ordination techniques (DCA) contained in the PC-ORD Version 3.0 statis- tical package (McCune and Mefford 1997). TWINSPAN (Hill 1979a) was used to identify floodplain forest types, and DCA (Hill 1979b) was used to illustrate the relationship between types. Default settings were used with the following exceptions: Braun- Blanquet cut-levels (0, 5, 26, 51, and 76) were used in the TWIN- SPAN analysis, and downweighting of rare species was used in DCA. Community types were based on both TWINSPAN and DCA results. Species indicator values for the community types were calculated using the Indicator Species Method of Dufréne and Legendre (1997) in the PC-ORD Version 3.0 statistical pack- age (McCune and Mefford 1997). Indicator species defined by the Indicator Species Method were used instead of TWINSPAN indicator species to describe community types because: (1) final community types were based on both TWINSPAN and DCA re- sults, and (2) the Indicator Species Method defines indicator spe- cies as those species present in the majority of sites belonging to a group, while TWINSPAN defines them as those species that are found mostly in a single group, but not necessarily in the majority of that group’s sites