Page  111 ï~~2003 THE MICHIGAN BOTANIST 111 COMPARISON OF CEDAR AND TAMARACK STANDS IN A RELICT CONIFER SWAMP AT PIERCE CEDAR CREEK INSTITUTE, BARRY COUNTY, MICHIGAN Bradford S. Slaughter' and J. Dan Skean, Jr. Department of Biology, Albion College Albion, MI 49224-1831 517.629.0525 dskean@albion.edu INTRODUCTION The Pierce Cedar Creek Institute is a relatively new environmental education and research center of ca. 258 hectares located in Barry County, in south-central lower Michigan. It encompasses parts of Sections 19 and 30 in Baltimore Township (T2N, R8W) in south-central Barry County, north and south of Cloverdale Road, roughly 12 km S of Hastings and 4 km W of highway M-37 at Dowling. The Institute is named for Cedar Creek, which flows north through the eastern part of the property. The area is notable for its interesting glacial topography and diversity of natural habitats. In addition to a prominent esker, the property includes Brewster Lake, a 3.5 hectare (9.5 acre) spring-fed kettle lake, which is known by local naturalists for its relatively pristine appearance and paucity of introduced species. The wetlands at the Institute include a plant community that would be classified by the Michigan Natural Features Inventory as relict conifer swamp (Chapman 1986). This swamp includes a conspicuous stand of northern white-cedar (Thuja occidentalis L.) that intergrades with tamarack (Larix laricina (DuRoi) K. Koch) along Cedar Creek. This naturally occurring "cedar swamp," which is likely the namesake for the creek, is probably one of Michigan's southernmost cedar swamps located outside of the lakeshore counties. Cedar swamps are widespread in northern Michigan, but are relatively rare in southern Michigan (Thompson 1953). Southern Michigan cedar swamps have been noted or studied by Cole (1901) and Wenger (1975) in Kent County, by Cole (1901) in Ottawa County, by Thompson (1953) in Oakland County, and by Pepoon (1927) in Cass County. These cedar swamps, with the exception of the Cass County site, are north of the Pierce Cedar Creek Institute location. Cedar swamps seem especially uncommon in southwest lower Michigan. Hanes & Hanes (1947) did not collect northern white-cedar in Kalamazoo County, and it was not recorded from Barry County or in surrounding counties in Voss (1972). Tamarack swamps, unlike cedar swamps, are relatively common throughout southern Michigan. Kost (2001a) noted that tamarack swamps 1Present address: Department of Botany, Miami University, Oxford, OH 45056 slaughbs@muohio.edu

Page  112 ï~~112 THE MICHIGAN BOTANIST Vol. 42 once covered nearly 200,000 hectares in southern lower Michigan, making them the most common type of conifer swamp in the southern part of the state. Tamarack, unlike northern white-cedar, was recorded in Barry County and in most surrounding counties by Voss (1972), and southern Michigan tamarackdominated wetlands have been studied by Kost (2001b), Kron (1989), Sytsma & Pippen (1982), Crow (1969a, 1969b), and Brewer (1966). In the early part of the last century, Transeau (1905, 1906) studied southern Michigan bogs in the Huron River valley where tamarack was present and sometimes occurred with black spruce (Picea mariana (Miller) BSP.), a species that is also more common in northern Michigan. The primary purpose of our study was to investigate the vascular flora of a southern Michigan cedar stand by comparing it to a nearby tamarack stand in the same relict conifer swamp. In doing so, our study focused on the possible consequences the presence of northern white-cedar has for vascular plant species diversity. A second purpose of this study was to assess the quality of the relict conifer swamp at the Pierce Cedar Institute by using measures developed by the Michigan Natural Features Inventory. Of interest to this study is a companion paper (Slaughter & Skean 2003), which is based on our general collecting in the area of our study sites. It presents an annotated checklist of the vascular plants in the vicinity of Cedar Creek and Brewster Lake, along with a rough vegetation map and brief descriptions of the major plant communities present. Because authors of species names are given in our other paper, they are omitted from the text below. PHYSICAL DESCRIPTION OF THE AREA The Pierce Cedar Creek Institute is in the Thornapple River drainage basin. The glacial topography of the area is most likely a result of ice stagnation of the Saginaw lobe during its final recession approximately 12,500 years ago (Roslund 2001). Over 80 hectares (200 acres) lie north of Cloverdale Road. This portion includes the entirety of Brewster Lake. Additionally, this section includes an esker, extensive wetlands, and areas of glacial till. The predominant upland soil in this area, according to Thoen (1990), is Marlette Loam, characterized by a layer of dark grayish brown loam resting on friable sandy loam, occurring on slopes, knolls, and ridges. Wetland soils are predominantly Houghton muck, a poorly drained soil characterized by a surface layer of black muck and a subsurface layer of black and dark reddish friable muck (Thoen 1990). The remaining portion of over 120 hectares (300 acres) lies south of Cloverdale Road, and includes nearly half of Section 30. This portion includes another esker, extensive wetlands, limited areas of till, and large areas of sand deposit in the western half (Roslund 2001). Several upland soils are present; esker soil is predominantly in the Marlette-Oshtemo complex, characterized by mixed Marlette (see above) and Oshtemo (dark brown sandy loam, often mixed with some clay) soils (Thoen 1990). Wetland soils are again predominantly mapped as Houghton

Page  113 ï~~2003 THE MICHIGAN BOTANIST 113 muck. Elevations range from around 258 meters in wetlands areas to over 288 meters in some upland sites (U. S. Geol. Survey 1982). Thoen (1990) also summarized average climactic conditions in Barry County. Average winter temperature, as recorded at Hastings, is -4.10C, and in summer is 20.80C. Average yearly precipitation is 79.3 cm, and average seasonal snowfall is 132 cm. RECENT HISTORY OF THE AREA According to the pre-European settlement vegetation map of Michigan (Comer and Albert 1998), the area that became the Pierce Cedar Creek Institute was dominated by oak-hickory forest W of Cedar Creek and beech-maple forest E of Cedar Creek, although wetlands, mapped S of Institute property, were certainly more extensive than indicated (see Discussion). European settlement began in the mid 1800's. Two Maryland brothers founded Baltimore Township in 1842 and established apple orchards; several other settlers followed suit shortly thereafter (Barry County Chamber of Commerce 2001). Baltimore Township retains a rural character today, and is dotted with farms and small settlements. It is the least populated township in Barry County, with fewer than 1900 inhabitants (Barry County Chamber of Commerce 2001). Township plat maps starting in 1928 show a variety of owners for the south parcel (NE Y4 S 30), which is the area in which the study sites were plotted. The 1928 plat map lists the owner as the "E. Mowry Estate," and the 1938 map lists Glen Mowry as the major property holder. In 1960, the property belonged to Orrie Dixon. The 1969 map lists H. Lewis Batts, Jr. as the parcel owner. The 1978 map lists Batts as the owner of the NW Y of Section 30 as well. A 1950 aerial photograph shows that, while most of the uplands surrounding the swamps were in cultivation or were disturbed by farming practices, the wetlands were essentially undisturbed. One exception is that part of Cedar Creek was ditched in the late 1950s or early 1960s, creating a large artificial wetland N of the tamarack swamp study site (Gary Pierce, pers. comm.). In 1998, the Willard G. Pierce and Jessie M. Pierce Foundation purchased the combined properties from H. Lewis Batts, Jr. The original purchase of 225 hectares is now the core area of the Pierce Cedar Creek Institute. The Institute provides ecological education to a variety of professionals and non-professionals, especially aimed at college students, and seeks to foster stewardship of the environment by a variety of means, including preservation, management, and restoration of natural landscapes (Pierce Cedar Creek Institute 2000). Much of the formerly farmed land has returned to forest, and people visit the Institute for its scenic beauty and its diversity of habitats and organisms. METHODS We compared the cedar and tamarack stands in terms of woody vegetation, herbaceous vegetation, soil pH, and standing water. We established study plots of 70 m by 70 m (4900 m2) in representative areas of both the cedar and tamarack stands in the NE 1/4 of Section 30 along the flood

Page  114 ï~~114 THE MICHIGAN BOTANIST Vol. 42 114 THE MICHIGAN BOTANIST Vol. 42 FIGURE 1. Locations of cedar stand and tamarack stand 4900 m2 study plots at the Pierce Cedar Creek Institute. plain of Cedar Creek (Figure 1) where the elevation in both sites ranges between 258 and 261 meters and the predominant soil is Houghton muck. We sampled eight N-S transects, spaced 10 m apart, utilizing stratified random sampling for both woody and herbaceous vegetation. We sampled woody vegetation (>2.5 cm diameter at breast height, d.b.h.) using the point-quarter method (Cottam and Curtis 1956) on September 11, 2001, and October 18, 2001. We determined the position of each sample point using a random numbers table. We took data at 27 locations in the cedar swamp, resulting in 108 trees sampled for distance from the point, species, and diameter at breast height, and 25 locations in the tamarack swamp, resulting in 100 trees sampled. From these data, we calculated relative density (RD), relative frequency (Rf), and relative coverage (RC) for each species. We calculated importance value (IV) for each species by adding these values, and assigned tree dominance based on importance values (Southwood and Henderson 2000). ANOVA and t-test statistical analyses were used in determining the size structure of tree species found in the swamps. On April 25, 2002, cores were taken with an increment corer from five cedars in the cedar plot and five tamaracks in the tamarack plot. Mostly larger trees were cored in order to estimate the ages of the stands. Methods for sampling herbaceous vegetation generally followed Kron (1989). Specifically, we determined locations of 1 m2 sample plots using a random numbers table. We surveyed four transects in May 2001 (May 18, 20, 23, 26, 29), and surveyed the other four transects in July 2001 (July 11, 25, 31). We sampled in these two months to study possible seasonal changes in herbaceous vegetation dominance, and took half our samples each day from each plot. We noted all species encoun

Page  115 ï~~2003 THE MICHIGAN BOTANIST 115 tered in each plot, and estimated their respective percent covers in each plot. Some sterile specimens that could not be identified reliably to species were grouped into larger taxa, e.g., Carex spp., Poaceae. We calculated relative cover (RC) and relative frequency (Rf) from the data. We then calculated importance values (IV) for each species by averaging relative cover and relative frequency (Kron 1989). Plants not identified in the field or not collected elsewhere on the property (see Slaughter and Skean 2003) were collected and deposited in the Albion College Herbarium (ALBC). In order to determine natural area quality of each site, we compiled detailed plant lists for each study plot, which included species not encountered in quantitative sampling, and assigned individual species coefficients of conservatism (C) and coefficients of wetness (W), as listed in Penskar et al. (2001) and described in Herman et al. (2001). We sampled soil pH in both the cedar and tamarack stands on October 30, 2001, using an Oakton Instruments pHTestr 3 waterproof pH meter. We took samples from topsoil cores at sites chosen using a random numbers table along transects from 12 points in the cedar swamp and 11 points in the tamarack swamp. Two samples were measured from each core. A two-tailed t-test was used to determine if there was a significant difference in pH between the swamp soils. Percent standing water was estimated for all lm2 sample plots taken in May and July. Again, a two-tailed t-test was used to determine if any significant differences in standing water existed between seasons and between study sites. RESULTS Cedar Stand Ten tree species were encountered in the cedar stand samples. Their importance values are listed in Table 1. The dominant tree was Thuja occidentalis (northern white-cedar), with an IV of 152. The five individuals of T occidentalis measured at d.b.h. and cored were the following sizes and ages: 8.1 cm (52 y), 22.1 cm (73 y), 29.5 cm (83 y), 37.6 cm (79 y), and 42.2 cm (85 y). Subdominant trees were Fraxinus nigra (black ash), IV = 60.8, and Betula alleghaniensis (yellow birch), IV = 32.0. Size class distributions for these three species are depicted in Figure 2. White-cedar trees showed a wide size class distribution, with diameters ranging from 5.3 cm to 36.8 cm. Nearly half of the white-cedars sampled were over 20 cm in diameter. ANOVA analysis indicated there was a difference in mean tree diameter between at least two species (F=26.3, df=2, TABLE 1. Results of the point-quarter sample of woody plants in the cedar stand at Pierce Cedar Creek Institute. Relative Relative Relative Importance Species Coverage Density Frequency Value Thuja occidentalis 75.5 44.4 31.9 152 Fraxinus nigra 10.7 24.1 26.1 60.8 Betula alleghaniensis 3.6 13.9 14.5 32.0 Larix laricina 6.9 5.6 8.7 21.1 Acer rubrum 0.4 4.6 7.2 12.3 Liriodendron tulipifera 2.1 2.8 4.3 9.2 Tilia americana 0.4 1.9 2.9 5.1 Carya cordiformis 0.4 0.9 1.4 2.8 Ulmus americana 0.1 0.9 1.4 2.4 Toxicodendron vernix 0.0 0.9 1.4 2.4

Page  116 ï~~116 THE MICHIGAN BOTANIST Vol. 42 18 16 0 Thuja occidentalis Fraxinus nigra 164 _ 14 - Betula alleghaniensis E 6 12 0 z 2.5-6 6-10 10-15 15-20 20-25 25-30 30+ Diameter at Breast Height (cm) FIGURE 2. Size class distributions for the three most important tree species in the cedar stand. P<0.001). T-tests showed individuals of both black ash (t=5.66, df=72, P<0.001) and yellow birch (t=5.38, df=61, P<0.001) were on average smaller than individuals of white-cedar. Diameters for black ash ranged from 2.5 cm to 25.4 cm, with over half of those individuals sampled being in the range of 3-6 cm. Species not sampled in the point-quarter method but found within plots for herbaceous species, as seedlings, included Acer saccharum (sugar maple) and Fagus grandifolia (beech), common species in the adjacent upland mesic southern forest. Liriodendron tulipifera (tuliptree) was also occasionally encountered in the cedar stand. A total of 55 herbaceous species was sampled and identified in the cedar stand. Table 2 lists the fifteen most important cedar stand herbaceous taxa for May and July with their respective importance values. The dominant herbaceous plant in the cedar stand was Symplocarpus foetidus (skunk-cabbage), with an importance value of 37.1 in May and 31.1 in July. It occurred in over 90% of the samples in both months. In addition, its large leaves translated to a very high relative percent cover. Subdominants included Mitella nuda (naked miterwort), Osmunda cinnamomea (cinnamon fern), Arisaema triphyllum (jack-in-the-pulpit), and various Carex spp. The cedar stand did not show a seasonal change in herbaceous plant composition; nine of the ten most important species in May were also among the ten most important species in July. One exception was Circaea alpina (dwarf enchanter's-nightshade), which was encountered at a much higher frequency in July (Rf = 2.97) than in May (Rf = 1.07). In total, 72 species of woody and herbaceous plants were identified in the cedar stand, 71 of them native. The only non-native species encountered was Taraxacum officinale (dandelion). The mean coefficient of conservatism (C) for cedar stand vegetation was 4.61, yielding a floristic quality index (FQ1) of 38.8. The average coefficient of wetness, or wetness index (W) for cedar stand plant

Page  117 ï~~2003 THE MICHIGAN BOTANIST 117 TABLE 2. Fifteen most important cedar stand herbaceous taxa for May and July. Relative Cover Relative Frequency Importance Value Taxon May July May July May July Symplocarpusfoetidus 67.5 55.7 6.70 6.53 37.1 31.1 Mitella nuda 3.87 3.45 5.09 5.64 4.48 4.54 Arisaema triphyllum 2.51 2.70 5.63 3.56 4.07 3.13 Osmunda cinnamomea 3.04 5.72 4.02 4.75 3.53 5.23 Maianthemum canadense 0.92 0.65 5.63 5.64 3.28 3.15 Carex spp. 2.36 4.42 3.22 3.56 2.79 3.99 Rubus pubescens 2.02 1.75 3.49 5.34 2.75 3.54 Galium aparine 0.66 1.04 4.83 4.75 2.75 2.89 Viola spp. 0.65 1.10 4.29 4.75 2.47 2.93 Impatiens capensis 2.78 * 2.14 * 2.46 * Trientalis borealis 0.69 * 3.75 * 2.22 * Cryptotaenia canadensis 1.30 1.94 2.41 3.96 1.86 2.75 Sterile Poaceae 0.65 0.76 2.95 3.26 1.80 2.01 Sterile Asteraceae 0.61 * 2.68 * 1.65 * Solidago rugosa 1.45 * 1.61 * 1.53 * Circaea alpina * 2.99 * 2.97 * 2.98 Apios americana * 2.23 * 2.37 * 2.30 Dryopteris intermedia * 1.46 * 2.97 * 2.21 Mitella diphylla * 3.36 * 0.89 * 2.13 * not among the 15 most important species for respective month species was -1.35 (FAC+). A complete listing of cedar stand species sampled and corresponding C and W values is found in Slaughter (2002, Table 3, with the following modifications: Cardamine pensylvanica and Polygonatum pubescens are now included, and Rorippa nasturtium-aquaticum and Uvularia grandiflora are omitted). Soil pH values ranged from 6.69 to 7.37, with a mean pH of 6.95. Standing TABLE 3. Results of the point-quarter sample of woody plants in the tamarack stand at Pierce Cedar Creek Institute. Relative Relative Relative Importance Species Coverage Density Frequency Value Larix laricina 62.6 24.0 19.7 106 Ulmus americana 19.9 27.0 23.9 70.9 Toxicodendron vernix 0.8 9.0 9.9 19.6 Betula alleghaniensis 3.5 7.0 8.5 19.0 Carpinus caroliniana 1.5 8.0 8.5 17.9 Cornus foemina 0.5 8.0 8.5 16.9 Thuja occidentalis 2.9 4.0 4.2 11.1 Quercus macrocarpa 3.5 3.0 4.2 10.8 Acer rubrum 3.2 3.0 4.2 10.4 Fraxinus nigra 0.7 4.0 4.2 8.9 Tilia americana 0.9 2.0 2.8 5.7 Prunus serotina 0.1 1.0 1.4 2.5

Page  118 ï~~118 THE MICHIGAN BOTANIST Vol. 42 water averaged 2.69% in May and 9.79% in July, with no statistically significant difference between the months (t=-1.34, df=48, P= 0.187). Tamarack Stand Twelve tree species were encountered in the tamarack stand samples. Their importance values are listed in Table 3. The dominant species was Larix laricina (tamarack), with an IV of 106. The five individuals of L. laricina measured at d.b.h. and cored were the following sizes and ages: 11.9 cm (45 y), 13.2 cm (35 y), 26.7 cm (60 y), 37.8 cm (48 y), and 40.1 cm (65 y). Subdominant trees were Ulmus americana (American elm), IV = 70.9, and Toxicodendron vernix (poison sumac), IV = 19.6. Size class distributions for these three species are depicted in Figure 3. Tamarack showed a wide size class distribution, with diameters ranging from 7.9 cm to 38.3 cm. Most sampled tamaracks were over 20 cm in diameter. ANOVA analysis again indicated there was a difference in mean tree diameter between at least two of the dominants (F=28.7, df=2, P<0.001). T-tests revealed tamaracks were, on average, larger than individuals of American elm (t=5.23, df=47, P<0.001) and poison sumac (t=6.20, df=30, P<0.001). Tree species found as seedlings in herbaceous plots but not encountered in pointquarter sampling included Fagus grandifolia (beech) and Carya cordiformis (bitternut hickory). Table 4 lists the fifteen most important tamarack stand herbaceous taxa for May and July with their respective importance values. The dominant herbaceous species in the tamarack stand was also skunk-cabbage, with an importance value of 33.2 in May and 18.4 in July. Subdominants included several unidentified sterile members of the family Poaceae (grasses), Impatiens capensis (spotted 10 9- O Larix laricina 2 2.5-6 6-10 10-15 15-20 20-25 25-30 30+ Diameter at Breast Height (cm)us americana FIGURE 3. Size class distributions for the three most important tree species in the tamarack stand. Â~ Toxicodendron vernix S6 M) 4 E 3-_ 2.5-6 6-10 ]10-15 ]15-20 20-25 25-30 30+ Diameter at Breast Height (cm) FIGURE 3. Size class distributions for the three most important tree species in the tamarack stand.

Page  119 ï~~2003 THE MICHIGAN BOTANIST 119 TABLE 4. Fifteen most important tamarack stand herbaceous taxa for May and July. Relative Cover Relative Frequency Importance Value Taxon May July May July May July Symplocarpusfoetidus 57.9 29.2 8.43 7.73 33.2 18.4 Sterile Poaceae 11.9 28.8 7.23 6.36 9.58 17.6 Carex spp. 6.31 1.22 4.82 2.27 5.56 1.75 Impatiens capensis 5.23 9.99 5.62 5.91 5.42 7.95 Carex lacustris 4.99 11.1 3.61 4.09 4.30 7.58 Lycopus uniflorus 2.61 1.61 5.22 4.55 3.92 3.08 Solidago rugosa 0.54 1.61 4.02 4.09 2.28 2.85 Equisetum fluviatile 0.11 * 4.42 * 2.27 * Thelypteris palustris 0.87 1.80 3.61 4.09 2.24 2.95 Sterile forb seedlings 2.78 * 1.61 * 2.19 * Galium aparine 0.14 0.44 4.02 4.09 2.08 2.27 Caltha palustris 0.77 * 2.81 * 1.79 * Boehmeria cylindrica 0.30 0.57 3.21 3.64 1.76 2.10 Mitella nuda 0.66 * 2.81 * 1.73 * Viola spp. 0.25 0.32 3.21 4.09 1.73 2.20 Aster lateriflorus * 1.58 * 4.09 * 2.83 Laportea canadensis * 2.61 * 1.36 * 1.98 Rubus pubescens * 0.48 * 3.18 * 1.83 Apios americana * 1.40 * 1.82 * 1.61 * not among the 15 most important species for respective month touch-me-not), several unidentified sterile members of the genus Carex (sedges), and Carex lacustris (common lakeshore sedge). The tamarack stand showed some seasonal differences; grasses (IV = 17.6) became co-dominant with skunkcabbage in July. Spotted touch-me-not and Aster lateriflorus (calico aster) were also more important in July than in May. A total of 73 herbaceous species was sampled and identified in the tamarack stand. In total, 85 species of woody and herbaceous plants were identified in the tamarack stand, all of them native. The mean coefficient of conservatism (C) for tamarack stand vegetation was 4.48, yielding a floristic quality index (FQ1) of 41.3. The wetness index (W) for tamarack stand plant species was -1.82 (FACW-). A complete listing of tamarack stand species sampled and corresponding C and W values is found in Slaughter (2002, Table 7, with the following modifications: Cardamine pensylvanica is now included, and Rorippa nasturtium-aquaticum and Uvularia grandiflora are omitted). Soil pH values ranged from 6.67 to 7.14, with a mean pH of 6.97. Standing water averaged 58.5% in May and 36.9% in July, with no statistically significant difference between the months (t=1.74, df=40, P=0.089). DISCUSSION Based on our sample, the cedar stand had a lower percentage of standing water than the tamarack stand. In the cedar stand, bare muck with no standing

Page  120 ï~~120 THE MICHIGAN BOTANIST Vol. 42 water was more commonly encountered than standing water. In May water cover averaged 2.7% in the cedar stand (n=26, s=9.62) and 58.5% in the tamarack stand (n=22, s=36.3), t=-7.55, df=46, P<0.001. In July water cover averaged 9.8% in the cedar stand (n=24, s=25.1) and 37% in the tamarack stand (n=20, s=44.2; t=-2.55, df=42, P=0.015). See Slaughter (2002, Table 4) for the complete data set. Although not conspicuous in the topographic map, the higher percentage of water cover in the tamarack stand might be due to a slightly lower elevation of this plot. There was no significant difference in soil pH between the cedar and tamarack stands (t=-0.63, df=44, P=0.529). The mean pH values for the cedar stand (6.95) and tamarack stand (6.97) fall within normal values of what is expected for relict conifer swamp. Chapman (1986) stated, "saturated muck soil is neutral due to groundwater infusion." Kron (1989) measured the pH in a tamarack swamp in Berrien County, Michigan, and obtained a value of 6.9. Southern Michigan relict conifer swamps studied by Kost (2001b), which were primarily tamarack swamps (none contained cedar), had soil pH ranging between 7.0 and 8.0. Relict conifer swamps are influenced by groundwater from calcium- and magnesium-rich glacial till that lies below the muck layer (Kost 2001b). The cedar stand is denser than the tamarack stand, which has a much more open appearance. This is reflected in the mean point-to-plant distances in our woody plant samples. The mean point-to-plant distance in the cedar stand (2.01 m) is significantly lower than that of the tamarack stand (3.08 m, t=-5.13, df=206, P<0.001). Within the cedar stand (Table 1), northern white-cedar appears especially dense, with a rather high relative density (44.4), compared to black ash (24.1), and yellow birch (13.9). In the cedar stand, there were many northern white-cedar trees in the larger size classes, and few in the smaller size classes (Figure 2). The largest cedars in our point-quarter samples approached 40 cm in diameter, not as large as the 60 cm diameter cedars Pepoon (1927) noted in a Cass County cedar swamp. When we searched out larger trees to core, we were surprised to find that the largest tree cored (ca. 42 cm d.b.h.) was 85 years old and that the smallest tree cored (ca. 8 cm d.b.h.) was 52 years old. An examination of the growth rings of the four large cedars (> 20 cm d.b.h.) indicates relatively rapid growth early on, followed by slower growth, perhaps as the forest canopy closed. The slower growth rate of the small cedar most likely reflects its position in the subcanopy. The sizes and ages of cedars in our small sample suggest that the cedars in the plot are less than a century old. In the cedar stand, no evidence of logging or catastrophic fire was apparent to us, but either could help explain the relative uniformity of the stand. Large mossy hummocks (including rotting logs) and adjacent mucky hollows characterize the ground layer of the cedar stand. Dead trunks and branches of Ulmus americana (American elm) were common on the swamp floor, indicating the cedar stand may have had a different tree composition before Dutch elm disease killed southern Michigan elms in the late 1950s (Catana 1967, Gary Pierce pers. comm.). Thompson (1953) suggested the presence of a cedar swamp in Oakland County may be explained by the protective influence of surrounding hills. In

Page  121 ï~~2003 THE MICHIGAN BOTANIST 121 deed, northern white-cedar is most prevalent at the Pierce Cedar Creek Institute in areas protected by an esker, as is the case for our cedar stand sample plot. As expected, L. laricina was the tree species with the highest importance value in the tamarack stand. However, more individuals of subdominant American elm (27) were sampled than tamarack (24). The high importance value of tamarack is due to the high number of individuals over 20 cm in diameter, including several trees over 30 cm that were sampled using the point-quarter method (Figure 3). The largest tamarack sought out for coring (ca. 40 cm d.b.h.) was also the oldest at 65 years. This suggests that the tamarack stand is even younger than the cedar stand. American elms were mostly in the 6-15 cm diameter range, and they inhabit the understory of the stand. In a study in Kalamazoo County, Michigan, Sytsma & Pippen (1982) found that size-structured canopies, dominated by large-diameter tamaracks, characterize old tamarack swamps. The Pierce Cedar Creek Institute tamarack stand sample plot fits this description, with numerous small hardwoods in the understory. A recent study of six conifer swamp sites by Kost (2001b) also found that tamarack dominates the upper size classes, while hardwoods dominate the smaller size classes. These results are also found in the cedar stand plot. Sytsma and Pippen (1982) suggest that succession of tamarack swamp to hardwoods is a slow, non-uniform process. Most likely, disturbances such as windthrow promote continued growth of tamaracks. An aerial photo of the study site taken in 1950 shows it to be a tamarack stand, similar to what is present today. Overall, the tamarack stand presents a mosaic appearance, with large trees, stumps, and associated mossy berms interspersed with open pools of standing water. Herbaceous flora of the berms is similar to that found in the cedar stand, but several different species are found in the pools of standing water, including many species typical of marshes or wet meadows (see below). The portion of the tamarack stand closest to Cedar Creek is dominated by young tamaracks and has a thick ground layer of grasses, sedges, and marsh species, suggesting that this area was once more open, and is now undergoing succession to tamarack swamp. Red maple (Acer rubrum) is occasionally found in both the cedar and the tamarack stands, but it is not a dominant or subdominant. Kost (2001b) implicates red maple as having the potential to negatively impact species richness in conifer swamp sites by causing succession to hardwood swamp. Five of six sites studied by Kost had red maple as a subdominant. Red maple is not this prevalent at the Pierce Cedar Creek Institute site, but it has the potential to become a subdominant in both stand types. Red maple was a minor component of presettlement forests, but it has become nearly ubiquitous in recent years, occupying a wide variety of communities (Abrams 1998). Skunk-cabbage was the most important herbaceous species in both the cedar and tamarack stands in May. In July it remained the most important herbaceous species in the cedar stand, and shared dominance with grasses in the tamarack stand. Of the ten most important cedar stand species for May, nine are springblooming shade dwellers. One exception is Impatiens capensis (spotted touchme-not), a summer and fall bloomer that is not uniform in distribution, but forms colonies in open areas. Of the ten most important July species, only Circaea alpina (dwarf enchanter's-nightshade) is a summer bloomer. The cedar stand

Page  122 ï~~122 THE MICHIGAN BOTANIST Vol. 42 herbaceous plants are mostly spring bloomers, but their vegetative parts remain green throughout the summer, unlike the spring ephemerals found in deciduous upland forests. Most of the species identified in the cedar stand are typical of other southern Michigan conifer swamps. Wenger (1975) found many of the same dominants in a Kent County, Michigan, cedar swamp. These include Maianthemum canadense (Canada mayflower), skunk-cabbage, jack-in-the-pulpit, cinnamon fern, and Viola spp. (violets). Brewer (1966), studying a bog forest in Kalamazoo County, Michigan, found several of these same species at high frequencies. Thompson (1953) noted that Mitella nuda (naked miterwort), C. alpina (dwarf enchanter's-nightshade) and Coptis trifolia (goldthread) were characteristic of an Oakland County cedar swamp. Emma Cole (1901) noted that, in the Grand Rapids, Michigan, region, naked miterwort was found locally in "deep cold woods," which happened to be cedar and tamarack swamps. Dwarf enchanter's-nightshade was found in "cold woods growing on or near decaying logs" (Cole 1901), a fitting description for its Pierce Cedar Creek Institute swamp habitat as well. Gaultheria hispidula (creeping snowberry), a sporadic species in southern Michigan, was found in one sample plot. It too shows a preference for conifer swamps (Voss 1996). The tamarack stand showed more seasonal variation in herbaceous species than the cedar stand (Table 4). In addition to grasses that shared dominance with skunk-cabbage in July, masses of Impatiens capensis (spotted touch-me-not) grew in importance from May to July. This is likely due to the more open, heterogeneous nature of the tamarack stand. Several large openings in the stand allow the growth of species that prefer openings and have summer blooming cycles. Several sedges are present in the tamarack swamp, including the subdominant Carex lacustris (common lakeshore sedge), which is common in open pools. Other species typical of open sites include Eupatorium maculatum (spotted Joe-Pye-weed), Eupatorium perfoliatum (boneset), and Cirsium muticum (swamp thistle). Species that thrive in the increased light provided by the tamarack stand have later blooming cycles than most species in the cedar stand, and thus become increasingly important. Our sample suggests that the tamarack stand has higher vascular plant species diversity than the cedar stand. A total of 85 species of vascular plants was identified in the tamarack stand plot, while 72 species were identified in the cedar stand plot. Fifty-seven of the species were shared between the plots. Fifteen species were exclusive to the cedar plot and 28 species were exclusive to the tamarack plot. In both plots, the actual number of herbaceous species encountered was probably higher due to the presence of sterile specimens that were not identified to species. This included several members of the genus Carex, as well as several sterile grasses and seedling forbs. The tamarack stand plot harbors 13 more identified species than the cedar stand plot, and this might be due to its more heterogeneous nature. Vivian-Smith (1997), in a study based on experimental wetland constructs, found that floristic diversity (species richness and evenness) was positively influenced by heterogeneous microtopography. Homogeneous habitats, on the other hand, provide optimal conditions for relatively few species, leading to competitive exclusion. Kost

Page  123 ï~~2003 THE MICHIGAN BOTANIST 123 (2001a) states that tamarack root mats form a varied microtopography that adds to the biocomplexity and high species richness of the tamarack stand community. The cedar stand study site is relatively heterogeneous, characterized by hummock-hollow topography, but it has little standing water and few large open areas. The tamarack stand has some hummock-hollow topography, but it also has flat open areas with standing water and small areas of drier ground. Kost (2001a) suggests the stark difference between moisture levels on the hummocks and in the saturated mudflats significantly increases the diversity of wetland ground flora. Several species, for example, were limited to a relatively dry microhabitat at the edge of the stand. These include Aquilegia canadensis (wild columbine), Asarum canadense (wild ginger), and Phlox divaricata (wood phlox). Mossy hummocks and rotting logs provided similar habitat to that found in the cedar stand, so most cedar stand species were also encountered in the tamarack stand. Many of these species are, however, more abundant in the cedar stand. Vivian-Smith (1997) found that many wetland species showed distinct preferences for either hummock or hollow microhabitats. Species that favored hummocks in our study, such as Coptis trifolia (goldthread), Mitella nuda (naked miterwort), Rubus pubescens (dwarf raspberry), and Trientalis borealis (starflower), were more frequent in the cedar stand, with its abundant hummock habitat, than in the tamarack stand, which does not show the same degree of hummock-hollow development. Gaultheria hispidula (creeping snowberry) was only encountered in a deep, mossy cedar stand hummock. Many species, such as skunk-cabbage and violets, showed variable microhabitat preferences and were more evenly distributed between both sites. Typha latifolia (cat-tail), Carex lacustris (common lakeshore sedge), and Boehmeria cylindrica (false nettle), which favor open, wet microhabitats, were only found in the tamarack stand. The tamarack stand appears to foster higher plant diversity due to its availability of a wide variety of microhabitats. The Floristic Quality Assessment (FQA) for the stands was influenced to some degree by our inability to identify certain sterile specimens. With this caveat, the values for the cedar stand (mean C = 4.61, FQI = 38.8) and tamarack stand (mean C = 4.48, FQI = 41.3) indicate that they are both high quality natural areas. According to Herman et al. (2001), areas with an FQI higher than 35 possess "sufficient conservatism and richness that they are floristically important from a statewide perspective." Studies of restored ecosystems, areas in which an attempt has been made to restore natural processes and species, have revealed that these projects rarely achieve mean C values of over 3.5 (Herman et al. 2001). Therefore, sites with values greater than 3.5 are judged to have at least marginal natural area quality, and those sites with values greater than 4.5 are almost certainly quality natural areas. Furthermore, the only introduced species encountered was the ubiquitous Taraxacum officinale (common dandelion), and this plant was not found in abundance. Inherent in the Floristic Quality Assessment is the fact that, as more species are identified, the FQI value will rise as long as the mean C value stays constant. Therefore, a more useful FQI for the region would include those species found outside of the sample plot and in surrounding communities. For this reason, the C value is more important in determining natural area quality in this case, since a FQI for the entire region, which would be appropriate, was not determined.

Page  124 ï~~124 THE MICHIGAN BOTANIST Vol. 42 The FQA does have some limitations. Several species found in the cedar stand are limited to swamps in southern Michigan, but are widespread in northern Michigan. For example, northern white-cedar, the dominant tree in the stand, is highly restricted in southern Michigan, and would justify a C value of 8 or 9. However, it occupies a wide variety of habitats in northern Michigan, and is assigned a C value of 4 for the state due to its geographically variable ability to colonize different habitats (Herman et al. 2001). Also, in theory, a community could be undisturbed but contain a preponderance of non-conservative native species. Furthermore, comparisons between different natural communities are perhaps best done based on C value, because there can be wide variations of species number between high quality communities of different types. For example, intact poor conifer swamps have lower diversity and therefore lower FQI values than similarly intact rich conifer swamps (Kost 2001b). The mean coefficient of wetness, or wetness index (WV), for the cedar stand (W = -1.35) and the tamarack stand (W = -1.82) indicate that both stands lie within the Facultative category, meaning the species are equally likely to occur in wetlands and in non-wetlands. Specifically, the cedar stand value falls under the category of Facultative + (FAC+), meaning the species are Facultative, but show a wet tendency (Herman et al. 2001). The W value is useful for wetlands delineation, but is not especially useful in categorizing the cedar stand vegetation. For example, wetland-restricted (W = -5) Osmunda regalis (royal fern) occurs in the cedar stand on mossy hummocks, but so does the rare specimen of upland-restricted (W = 5) Trillium grandiflorum (large white trillium). The wetness index is more useful as an overall figure, because it tends to marginalize the fact that microhabitats within wetlands often allow individual "restricted" species of both uplands and wetlands to coexist, often on the same mossy hummock. The wetness index (WI) for the tamarack stand falls under the category of FACW-, or Facultative Wetland - (Herman et al. 2001). Plants falling under this category are said to occur in wetlands 67-99% of the time; the "minus" value indicates a dry tendency, so the 67% value is probably a more reliable number. This value suggests plant species found in the tamarack stand, on average, have a greater tendency towards growing in wetlands than plant species found in the cedar stand. This is to be expected, as the tamarack stand has less upland microhabitat and much more standing water than the cedar stand. The values should not be misinterpreted to mean that the tamarack stand is "wetter" than the cedar stand, however. The index simply measures the probability a given plant species will occur in a wetland. A dry, mossy hummock and a marsh with standing water may be equally likely to harbor an Obligate Wetland species. The value does suggest that there may be greater availability of upland microhabitat in the cedar stand than in the tamarack stand. The specific area of both sample plots is indicated as oak-hickory forest on the pre-settlement vegetation map of Michigan (Comer and Albert 1998), but rather extensive mixed conifer swamp is mapped further S along Cedar Creek. Aerial photographs, soils, hydrology, and vegetation, however, suggest the wetlands along Cedar Creek were much more extensive than mapped. Comer and Albert (1998) indicate that survey lines often missed localized habitat types, and that some specific habitats located in complex landscapes were not indicated by

Page  125 ï~~2003 THE MICHIGAN BOTANIST 125 the surveys. The relict conifer swamps at Pierce Cedar Creek Institute likely fall under one of those two categories. Most relict conifer swamp communities mapped in southern Michigan are now interpreted to represent tamarack swamps (Kost 2001a). The cedar and tamarack sample plots are in the same drainage basin, and can both be classified as relict conifer swamp based on tree importance values. However, the two study sites showed considerable differences in dominant trees, tree density, herbaceous vegetation, and percent standing water. Because most plant species were found in both the cedar and tamarack stands, the presence of northern white-cedar is most likely not a major factor in determining species diversity. In fact, more species were encountered in the tamarack stand, a more heterogeneous habitat than the cedar stand. Our results suggest that there can be considerable differences in species distribution based on the availability of microhabitats suitable for plant colonization, even though the entire area may be classified as the same community type. Further studies should be performed on swamps west of the esker to determine an overall species list for this community type. Also, a more detailed Floristic Quality Assessment (FQA) should be performed for the swamps and bordering mesic southern forest. This study should take into account the numerous species found in edge habitats, which harbor many species due to the convergence of numerous community types. In any event, our work suggests that the Pierce Cedar Creek Institute harbors high-quality natural habitats that are worthy of continued protection and study. ACKNOWLEDGMENTS We thank Gary J. Pierce, Director of the Pierce Cedar Creek Institute for his cooperation and support. We also thank Dean G. McCurdy, Assistant Professor of Biology, and Hal H. Wyss, Professor of English, both at Albion College, for their comments on an earlier draft of this paper. The first author thanks the benefactors of the Ewell A. and Barbara J. Stowell Scholarship in Environmental Biology, and the Marilyn Young Vitek Merit Scholarship in Biology. The second author thanks the benefactors of the A. Merton Chickering Professorship at Albion College. The Willard G. Pierce and Jessie M. Pierce Foundation and the Albion College Foundation for Undergraduate Research, Scholarship, and Creative Activity (FURSCA) provided additional support. This paper is based on a thesis submitted by the first author in partial fulfillment of requirements for the Bachelor of Arts with Departmental Honors in Biology at Albion College. LITERATURE CITED Abrams, M. D. 1998. The red maple paradox: What explains the widespread expansion of red maple in eastern forests? BioScience 48: 355-364. Barry County Chamber of Commerce. 2001. Baltimore Township: The most unexpected township. Privately published brochure. Brewer, R. 1966. Vegetation of two bogs in southwestern Michigan. Michigan Botanist 5: 36-46. Catana, A. J., Jr. 1967. Forests of the Harvey N. Ott Preserve. American Midland Naturalist 78: 496-507. Chapman, K. A. 1986. Michigan Natural Community Types. Michigan Natural Features Inventory. <http://www.dnr.state.mi.us/wildlife/heritage/mnfi/lists/ natural_community_types.pdf> (Accessed 19 March 2002). Cole, E. J. 1901. A Grand Rapids Flora: A Catalogue of the Flowering Plants and Ferns Growing Without Cultivation in the Vicinity of Grand Rapids, Michigan. A. Van Dort, Grand Rapids.

Page  126 ï~~126 THE MICHIGAN BOTANIST Vol. 42 Comer, P. J. & D. A. Albert. 1998. Presettlement vegetation of Michigan: An interpretation of the General Land office surveys. Michigan Natural Features Inventory, Lansing. 2 maps. Cottam, G. & J. T. Curtis. 1956. The use of distance measures in phytosociological sampling. Ecology 37: 451-460. Crow, G. E. 1969a. An ecological analysis of a southern Michigan bog. Michigan Botanist 8:11-27. Crow, G. E. 1969b. Species of vascular plants of Pennfield Bog, Calhoun County, Michigan. Michigan Botanist 8: 131-136. Hanes, C. R. & F. N. Hanes. 1947. Flora of Kalamazoo County, Michigan. Vascular Plants. [Anthoensen Press; Portland, Maine.], Schoolcraft, Michigan. Herman, K. D., Masters, L. A., Penskar, M. R., Reznicek, A. A., Wilhelm, G. S., Brodovich, W. W., & K. P. Gardiner. 2001. Floristic Quality Assessment with wetland categories and examples of computer applications for the State of Michigan. Revised, 2nd Edition. Michigan Department of Natural Resources, Wildlife, Natural Heritage Program. Lansing. 19 pp. + Appendices. Kost, M. A. 2001a. Natural community abstract for relict conifer swamp. Michigan Natural Features Inventory, Lansing. 6 pp. Kost, M. A. 2001b. Potential indicators for assessing biological integrity of forested, depressional wetlands in southern Michigan. Michigan Natural Features Inventory, Lansing. 69 pp. Kron, K. A. 1989. The vegetation of Indian Bowl wet prairie and its adjacent plant communities. I. Description of the vegetation. Michigan Botanist 28: 179-200. Penskar, M. R., Reznicek, A. A., Brodovich, W. W., Wilhelm, G. S., Masters, L. A., Herman, K. D., and K. P. Gardiner. 2001. Michigan plants database. In Herman, K. D., Masters, L. A., Penskar, M. R., Reznicek, A. A., Wilhelm, G. S., Brodovich, W. W., & K. P. Gardiner. 2001. Floristic Quality Assessment with wetland categories and examples of computer applications for the State of Michigan. Michigan Department of Natural Resources, Wildlife Division, Natural Heritage Program, Lansing. 19 pp. + Appendices. Pepoon, H. S. 1927. Flora of the Chicago Region. Lakeside Press, Chicago. Pierce Cedar Creek Institute. 2000. Welcome to the Pierce Cedar Creek Institute. Privately reproduced brochure. Roslund, K. K. 2001. Field and GIS mapping of surficial glacial deposits and landforms near Dowling, Michigan. Undergraduate Honors Thesis, Albion College, Michigan. Slaughter, B. S. 2002. Comparison of a cedar swamp and a tamarack swamp at Pierce Cedar Creek Institute, Barry County, Michigan, with an annotated checklist of vascular plant species. Undergraduate Biology Honors Thesis, Albion College, Albion, Michigan. ____ and J. D. Skean, Jr. 2003. Annotated checklist of vascular plants in the vicinity of Cedar Creek and Brewster Lake, Pierce Cedar Creek Institute, Barry County, Michigan. Michigan Botanist 42: 129-150. Southwood, T. R. E. & P. A. Henderson. 2000. Ecological methods. Blackwell, London. Sytsma, K. J. & R. W. Pippen. 1982. The Hampton Creek wetland complex in southwestern Michigan. III. Structure and succession of tamarack forests. Michigan Botanist 21: 67-74. Thoen, G. F. 1990. Soil survey of Barry County, Michigan. United States Department of Agriculture. Washington, D.C. Thompson, P. W. 1953. Vegetation of Haven Hill, Michigan. American Midland Naturalist 50: 218-223. Transeau, E. N. 1905. Th e bogs and bog flora of the Huron River Valley. Botanical Gazette 40: 351-375, 418-448. Transeau, E. N. 1906. The bogs and bog flora of the Huron River Valley [concluded]. Botanical Gazette 41: 17-42. U.S. Geological Survey. 1982. Dowling, Michigan Quadrangle. U.S. Geological Survey. Reston, Virginia. Provisional Map. Vivian-Smith, G. 1997. Microtopographic heterogeneity and floristic diversity in experimental wetland communities. Journal of Ecology 85: 71-82. Voss, E. G. 1972. Michigan flora. Part I. Gymnosperms and monocots. Bull. Cranbrook Inst. Sci. 55 and Univ. Michigan Herbarium, Ann Arbor, Michigan. Voss, E. G. 1996. Michigan flora. Part III. Dicots (Pyrolaceae-Compositae). Bull. Cranbrook Inst. Sci. 61 and Univ. Michigan Herbarium, Ann Arbor, Michigan. Wenger, J. D. 1975. The vegetation of a white-cedar swamp in southwestern Michigan. Michigan Botanist 14: 124-130.