Methods 



some 1500 people, whose lives cen- 

 tered around a massive granite quarry- 

 ing operation. When the market for 

 water-transported granite collapsed 

 in the early 1900's, the entire populace 

 abandoned the island. The effects of 

 such extensive human use on island 

 topography and vegetation are still 

 apparent. A detailed history of the 

 Penobscot Bay islands is recounted 

 by McLane (1982). 



The maritime climate of the 

 Penobscot Bay islands has been de- 

 scribed by Davis (1966) and Conl<ling 

 (1981). It is affected by the Nova Scotia 

 Current, which draws arctic waters 

 into the Gulf of Maine with the effect 

 of lowering summer temperatures 4° 

 to 6 °C below those at inland stations. 

 Warm, moisture-laden air from the 

 offshore Gulf Stream passing over 

 this cooler current produces extended 

 periods of fog early in the growing 

 season. Thornthwaite (1948) classified 

 the Maine coast as "perhumid", the 

 most humid category of North Ameri- 

 can regions, in addition, heat storage 

 in the waters of the Gulf of Maine 

 lengthens the growing season in the 

 autumn. This extended, relatively cool 

 and very humid growing season yields 

 a transitional flora with both temper- 

 ate and boreal affinities (Hill 1923). 



A dense spruce-fir forest covers 

 most of Hurricane Island. The forest 

 floor is primarily composed of mor, 

 mosses, and lichens. The understory 

 is sparse, except where the canopy 

 is relatively open or soils are too 

 shallow to support an overstory. Domi- 

 nant groundcover species include 

 such rhizomateous perennials as Mai- 

 anthemum canadense (Canada may- 

 flower), Linnaea borealis (twinf lower), 

 Clintonia borealis (bluebead-lily), 

 Cornus canadensis (bunchberry), and 

 Aralia nudicaulis (sarsaparilla). Char- 

 acteristic shrubs include Vaccinium 

 augustifolium (lowbush blueberry), 

 Gaylussacia bactaca (huckleberry), 

 Myrica pensylvanica (bayberry), and 

 Juniperus liorizontalis (horizontal 

 juniper). 



During the summers of 1980 and 

 1981, 15 simulated trails (labeled E 

 through S) were established at loca- 

 tions that showed no signs of recent 

 human disturbance. Trail sites were 

 chosen to represent distinct species 

 or plant associations. Dominant 

 species at each trail are listed in 

 Table 1. References for nomenclature 

 were as follows: flowering plants: 

 Gray (Fernald 1950); mosses: Crum 

 (1973) and lichens: Hale (1979). Two 

 to six study plots, measuring 1 x 

 0.5 m were located on each trail. Plots 

 were placed at locations with repre- 

 sentative samples of selected vegeta- 

 tion and were not necessarily adjacent 

 to each other on the trails. Each study 

 plot was marked by a pair of stakes 

 positioned 1 m apart on opposite 

 sides of the trail. One stake of each 

 pair was marked with a plot-identifying 

 code. If soil depth was inadequate to 

 support a stake, the corners of the 

 plot were marked with red paint. 

 Where no rock was exposed for paint 

 marking, a tape was used to measure 

 between tags affixed to roots or trees 

 to mark the exact location of the plot. 



The quadropod— a 4-legged in- 

 strument with a 1- X 0.5-m rectangular 

 base frame which delineates the study 

 plot— was used to record ground- 

 cover changes over the study period. 

 Photographs of the plot were taken by 

 a camera mounted at the quadropod's 

 apex and aimed directly downward so 

 that the entire frame was in view. 

 Variations in photographic technique 

 are minimized, since the quadropod 

 holds the film plane parallel with 

 the ground surface, maintains a con- 

 stant distance from the lens to the 

 ground, and reduces vibrations. A 

 moredetailed description of thequad- 

 ropod system is given by McBride and 

 Leonard (1982). 



Trails E through L were trampled 

 a total of 100 times in three incre- 

 ments of 40, 40, and 20 times in July 

 and August of 1980. Plots were photo- 

 graphed at 0, 80, and 100 tramples. 

 Trails H-2 and M through S were tram- 

 pled a total of 200 times in increments 

 of 0, 80, and 100 tramples in July and 



August of 1981, and 150 and 200 

 tramples in July and August of 1982. 

 Plots were photographed 2 weeks 

 after each trampling. Trampling dam- 

 age was measured more accurately 

 2 weeks after trampling than on the 

 same day. Two field assistants, aver- 

 aging 135 pounds (61.2 kg), walked 

 over the simulated trails a specified 

 number of times on each trampling 

 date. They wore soft-soled footwear 

 instead of lug-soled boots because 

 that is the typical footwear of recre- 

 ationists coming to islands by boat. 

 One year of recovery data were ob- 

 tained for Trails M through S (1983) 

 and 2 years of recovery data for Trails 

 E through L (1981, 1982). Plots were 

 photographed two or three times over 

 the growing season of each recovery 

 year. Photographs were analyzed in 

 the laboratory. Projected images of 

 the plot were mapped. Each plot was 

 divided into three sections: a central 

 0.5- X 0.5-m section (Area 1) and two 

 .25- X 0.5-m side sections (grouped 

 together as Area 2). Trampling oc- 

 curred over the center portion of 

 Area 1. Area 2 of each plot was desig- 

 nated as a control. Percent coverages 

 of species in Areas 1 and 2 were 

 measured with a planimeter or by 

 counting individual plants. 



Relative percentages for each 

 species were obtained in the follow- 

 ing manner for analysis: 



Relative percent coverage = 

 percent absolute coverage 



greatest percent coverage 

 during the trampling phase 



where: 



— percent absolute coverage repre- 

 sents the area of the study plot 

 which is covered by the plant spe- 

 cies on a given measurement date 



—greatest percent coverage during 

 the trampling phase represents the 

 greatest area coverage measure- 

 ment of that plant species taken 

 at any time during one trampling 

 phase or season. 



2 



