Page  118 ï~~118 THE MICHIGAN BOTANIST Vol. 49 TREE COMPOSITION AND DEMOGRAPHY OF A SECOND GROWTH HARDWOOD FOREST IN BERRIEN COUNTY, MICHIGAN Robert Tatina Dakota Wesleyan University Biology Department Mitchell, SD 57301 rotatina@dwu.edu ABSTRACT The tree species composition of a small, second growth forest in Berrien County, Michigan, about one-half mile east of Lake Michigan, was determined from tree numbers and diameters gathered using the T-square method. The forest is dominated by northern red oak (Quercus rubra), eastern white pine (Pinus strobus) and red maple (Acer rubrum). When compared to upland sites, lowland sites had greater species diversity (14 compared to 11) and a greater basal area (57.84 compared to 39.16 m2. ha-1), but had a lower density (317.89 compared to 357.05 trees - ha-1) of trees. Size class structure of the tree species generally showed that they were replacing themselves and will continue for some time into the future. KEY WORDS: Berrien County, Michigan, Southern Mesic Hardwood Forest, Species Composition, Tree Size Classes INTRODUCTION According to original land survey records, the vegetation of Berrien County, Michigan, prior to settlement by non-native people was hardwood forest with dominance shared by American beech (Fagus grandifolia) and sugar maple (Acer saccharum) (Brewer et al. 1984). Since then much of the county has been logged (Coolidge 1906) and some of it burned (Ellis 1880), leaving a mosaic of agricultural land, forests and commercial and residential developments. Over time, the once grand forests of Michigan have become reduced in area, and today represent about half of their former extent (Dickmann and Leefers 2003). Presently, most of the forested stands are second growth. Two forests in Berrien County have been the subject of continuing research. These are the beech-maple forest at Warren Woods State Park studied by Billington (1924), Cain (1935), Brewer and Merritt (1978), Woods (1979), Brewer (1980), and Donnelly and Murphy (1987) and the several forest types at Warren Dunes State Park described by Wells and Thompson (1982), Smith and Woodland (2006) and Smith and Woodland (2007). These have provided much insight into the composition and dynamics of old growth, hardwood forests as well as a benchmark against which to compare other forests. However, most of the forests in southwestern Michigan, as well as the rest of the state, are second growth and young with little attention paid to their composition and dynamics.

Page  119 ï~~2010 THE MICHIGAN BOTANIST 119 This study was conducted in a small, second growth forest in southwestern Michigan about five miles north of Warren Woods State Park and within a half mile south of the woods at Warren Dunes State Park. Prior to sampling the trees, I compiled a list of 184 vascular plant species and from it estimated the Floristic Quality Index to be greater than 50. My objectives were 1) to describe the composition of the forest in terms of its tree species and 2) to examine the size class (age) structure of the tree species to estimate what the composition of this forest might be in the future. STUDY AREA The study area (Figure 1) is a forested stand of approximately 8.1 ha, ~ 1 km WNW of Sawyer, Michigan. Its legal description is T7S, R20E, Sec. 2, NW 1/4, SW 1/4. On the south it is bounded by Lake Lane and on the west by Tower Hill Road. From the southwest corner, the area extends -250 m N and -400 m E. It is the southern end of the Tower Hill Camp and Retreat Center (THC), a 22-hectare tract which was donated to the Illinois Conference of the United Church of Christ by Edward K. Warren in 1923 (Gregg Briggs, THC Director, personal communication). The site is approximately 0.75 km southeast of Lake Michigan. Although I was unable to find specific information about the property prior to it being converted into a camp, circumstantial evidence suggests that the area was logged. In 1854 a saw mill was operating in Sawyer, and in 1861 there was another saw mill operating in the vicinity with a horse drawn railroad running through the area of the camp down to a pier on Lake Michigan at the Chikaming-Lake Township boundary (Coolidge 1906). These historical notes would suggest that the area had been logged as early as the mid 1800s. The age of the stand, estimated from tree cores of some of the larger red oaks, is about 70-80 years old. Immediately to the north of the stand is a white pine planting with trees standing in obvious rows. These trees were aged at 70-80 years also. The absence of sawed tree stumps indicates that the area had not been logged recently. In addition, there is no evidence of recent fire. The topography is rolling with small hills that peak at approximately 200 m above sea level and low areas at 190 m. The soils are sandy and have been mapped as Oakville fine sands on ridge tops, knolls and slopes and Pipestone sand and Grandby loamy fine sand on nearly level areas (Larson 1980). The Oakville fine sands have rapid infiltration rates, and the Pipestone sand and Grandby loamy fine sand are poorly drained because of a high water table (Larson 1980). Generally, the forest fits a southern mesic hardwood forest as described by Kost et al. (2007). METHODS During the summer, 2009, starting at the SW corner of the study area, I laid out a grid of sampling points 20 m in from the property lines and 20 m apart. From the resulting 250 intersecting lines, -150 intersections were chosen at random to become sampling point centers. (Excluded from sam

Page  120 ï~~120 THE MICHIGAN BOTANIST Vol. 49 pling was a portion of the woods where white pines had been planted in obvious rows and marked as "Pines" in Figure 1.) To determine which two trees at each sampling point were to be measured, I followed the T-square plotless sampling method because it overcomes sampling biases when trees are not randomly distributed (Greenwood 1996). (I chose this method as an alternative to determining the actual distributional pattern of the trees.) Selecting the tree nearest a sampling point, I measured the distance between its center and the sampling point and then measured the circumference of the tree at breast height (1.4 m). Following the sampling procedures of Curtis (1959), I designated tree size class individuals as those that had a diameter at breast height (dbh) 10 cm. The second tree to be measured was the nearest neighbor to the first tree found beyond a line drawn through the first tree perpendicular to a line extending from the sampling point to the first tree. I then measured the distance between the first tree and the second tree, and measured the circumference of the second tree at breast height. Saplings (tree species whose dbh < 10 cm and whose height 30 cm) were counted by dbh size class (< 2; 2 <4; > 4 < 6; > 6 < 8; > 8 < 10 cm) in 0.01 ha circular plots (radius = 5.64 m) whose radii extended from each sampling point. Seedlings (< 2 cm dbh and < 30 cm tall) of tree species were counted in 0.002 ha circular plots (radius = 2.52 m) centered on the sampling point. I also noted whether the point and the sampling area were located in a lowland or an upland. Generally, the lowland points were located at the southern and eastern edges of the forest and had water standing in depressions in the spring, 2009. The uplands were on shallow slopes and hills and showed no evidence of having been inundated. Density of tree size class species was determined from distance measurements using the following: density = n2/2.28281xly, in which n is the number of sampling points, x is the distance to the nearest tree from a sampling point and y is the distance from the first tree to its nearest neighbor (Greenwood 1996). For sapling and seedling species, density was calculated by dividing the number of individuals counted by the total area sampled. Tree circumferences were converted into basal areas and then summed by species; points of occurrence were converted into frequency by dividing the number of points of occurrence by the number of points sampled. Density, basal area and frequency for each tree size class species were relativized and then summed to produce importance values (IV). With an increment borer I extracted cores at breast height from several larger trees. After drying the cores I counted the annual rings to determine the approximate age of each tree. Scientific nomenclature follows that of Gleason and Cronquist (1991), and common names of trees that of Barnes and Wagner (2004). RESULTS Increment cores taken from seven species of larger trees produced ages that fell into two groups (Table 1). Except for Fagus grandifolia, which averaged about 125 years old, all of the other species were 70 to 80 years old. Table 2 shows the tree species composition for the entire forest. The stand had a total density of nearly 342 trees " ha-1 and a total basal area of 122.50 m2. ha-1. The average distance between trees was 3.95 m Â~ 2.39 (sd). Based on IVs, dominant tree species were Quercus rubra (northern red oak), Pinus strobus (eastern white pine) and Acer rubrum (northern red maple). Acer rubrum existed at the highest density (69.05 trees - ha-1). At an average dbh of 0.630 m, Liriodendron tulipifera (tulip tree) comprised the largest trees, one of these having a dbh of 1.251 m. Other common tree species included Prunus serotina (black cherry), Fagus grandifolia, Nyssa sylvatica (black gum) and Acer saccharum (sugar maple). Tables 3 and 4 show the composition of the forest when sampling points are divided into lowlands and uplands. The 91 trees sampled at 42 points in the

Page  121 ï~~2010 THE MICHIGAN BOTANIST 121 2010 THE MICHIGAN BOTANIST 121 FIGURE 1. Aerial view of Tower Hill Camp and Retreat Center, Berrien County, Sawyer, Michigan. The areas sampled in this study are indicated as Upland and Lowland. Excluded was the area labeled "Pines". lower, wetter portions of the study area were divided among 14 species which existed at a density of 317.89 trees " ha-1 (Table 3). The average distance between trees was 4.24 m Â~ 2.54 (sd). On the lowland sites Acer rubrum and Liriodendron tulipifera were the most important trees species, with Acer rubrum having the highest density (96.74 trees " ha-1) and greatest basal area per hectare

Page  122 ï~~122 THE MICHIGAN BOTANIST Vol. 49 TABLE 1. Ages and diameters (dbh) of tree species from Tower Hill Camp and Retreat Center, Sawyer, Michigan. Where more than one tree was cored, values for ages and diameters are means Â~ sd. SPECIES AGE (years) dbh (meters) N Acer rubrum 73.00 Â~ 20.25 0.65 Â~ 0.15 4 Pinus strobus 72.83 Â~ 4.67 0.53 Â~ 0.22 6 Fagus grandifolia 125.50 Â~ 2.12 0.73 Â~ 0.13 2 Prunus serotina 81 0.79 1 Liriodendron tulipifera 85 0.55 1 Sassafras albidum 88 0.46 1 Picea abies 82.00 Â~ 2.00 0.51 Â~ 0.19 3 (21.87 m2. ha-1). Other frequently encountered species were Prunus serotina, Quercus rubra and Nyssa sylvatica. Species encountered at lowland sites, but not upland sites were Betula alleghaniensis (yellow birch), Picea abies (Norway spruce), Carpinus caroliniana (blue-beech), Carya ovata (shagbark hickory) and Fraxinus pennsylvanica (red ash). The 81 points sampled at upland locations contained 11 species at a total density of 357.05 trees " ha-1 (Table 4). The average distance between trees was 3.82 m Â~ 2.28 (sd). Pinus strobus and Quercus rubra shared dominance here and together accounted for 63% of the trees. Two species, Ostrya virginiana (hophornbeam) and Quercus alba (white oak) were found at upland points only, albeit infrequently. When compared to upland sites (Table 4), lowland sites (Table 3) were found to have a greater species diversity (14 compared to 11) and a greater basal area (57.84 compared to 39.16 m2. ha-1) but a lower density of trees (317.89 compared to 357.05 trees " ha-1). The average distance between trees at lowland sites was not significantly different from that of the uplands (t = 0.91, df = 85, P = 0.37). Although Acer rubrum and Liriodendron tulipifera were present at both lowland and upland sites, these two species had higher IVs at lowland sites. The opposite statement can be made for Pinus strobus and Quercus rubra whose IVs were significantly higher at upland sites. Figure 2 shows the density of tree species by one seedling, five sapling and two tree size classes. For Acer rubrum, Fagus grandifolia (Figure 2A, B), Acer saccharum (Figure 2C), Prunus serotina (Figure 2D) and Nyssa sylvatica (Figure 2E) there was a general decrease in density from seedling through tree size classes. In most of these species, the 25-cm dbh midpoint size class showed an increase over the larger sapling size classes. For example, in Acer rubrum (Figure 2A) the 25 cm size class exceeded the 7 cm and 9 cm size classes. This anomaly may be an artifact of having divided all of the stems greater than 10 cm dbh into only two size classes and might disappear if they had been divided among more size classes. While not abundant in any size class, Fagus grandifolia (Figure 2B), Acer saccharum (Figure 2C) and Nyssa sylvatica (Figure 2E) had low densities of trees larger than 25 cm. Quercus rubra stem densities (Figure 2F) generally

Page  123 ï~~2010 THE MICHIGAN BOTANIST 123 TABLE 2. Tree species composition for all grid points in woods at Tower Hill Camp and Retreat Center, Sawyer. Michigan. ACRU = Acer rubrum, ACSA = Acer saccharum, BEAL = Betula alleghaniensis, CACA = Carpinus caoliniana, CAOV = Carya ovata, FAGR = Fagus grandifolia, FRPE = Fraxinus pennsylvanica, LITU = Liriodendron tulipifera, NYSY = Nyssa sylvatica, OSVI = Ostrya virginiana, PIAB = Picea abies, PIST = Pinus strobus, PRSE = Prunus serotina, QUAL = Quercus alba, QURU = Quercus rubra, SAAL = Sassafras albidum, # Trees = number of trees counted and measured, BA = basal area, Rel Dom = Relative Dominance, Rel Dens = Relative Density, # Pts = Number of grid points of occurrence, Rel Freq = Relative Frequency, IV = Importance Value (Sum of Rel Dom, Rel Dens and Rel Freq). Mean BA # (sd) Max (m2/ Rel Trees/ Rel Rel Spp. Trees dbh dbh ha) Dom ha Dens # Pts Freq Freq IV QURU 57 0.361 1.152 24.920 8 69.05 20.22 40 0.284 18.78 60.19 (0.219) PIST 62 0.302 1.069 22.680 12.85 75.10 21.99 40 0.284 18.78 53.62 (0.163) ACRU 43 0.415 1.075 21.613 20.37 52.08 15.25 35 0.248 16.43 52.05 (0.235) LITU 23 0.630 1.251 17.552 21.61 27.86 8.16 20 0.142 9.39 39.16 (0.232) PRSE 26 0.346 0.579 10.899 7.56 31.5 9.22 23 0.163 10.80 27.58 (0.141) FAGR 19 0.345 0.793 7.942 6.54 23.02 6.74 14 0.099 6.57 19.85 (0.219) NYSY 16 0.249 0.522 4.826 2.53 19.38 5.67 12 0.085 5.63 13.83 (0.120) ACSA 11 0.236 0.057 3.143 1.91 13.32 3.90 9 0.064 4.22 10.03 (0.173) SAAL 9 0.247 0.360 2.690 1.24 10.90 3.19 8 0.057 3.75 8.18 (0.074) PIAB 6 0.408 0.541 2.966 2.19 7.27 2.13 4 0.028 1.88 6.20 (0.098) BEAL 3 0.281 0.446 1.020 0.58 3.63 1.06 2 0.014 0.94 2.58 (.0143) QUAL 2 0.404 0.433 0.978 0.69 2.42 0.71 1 0.007 0.47 1.87 (0.040) OSVI 2 0.170 0.194 0.411 0.12 2.42 0.71 2 0.014 0.94 1.77 (0.035) FRPE 1 0.458 0.458 0.554 0.56 1.21 0.35 1 0.007 0.47 1.38 CAOV 1 0.134 0.134 0.162 0.04 1.21 0.35 1 0.007 0.47 0.86 CACA 1 0.120 0.120 0.145 0.03 1.21 0.35 1 0.007 0.47 0.85 Totals 282 122.501 86.82 341.58 100 213 1.511 99.99 300.00

Page  124 ï~~124 THE MICHIGAN BOTANIST Vol. 49 TABLE 3. Tree species composition for lowland grid points in woods at Tower Hill Camp and Retreat Center, Sawyer. Michigan. ACRU = Acer rubrum, ACSA = Acer saccharum, BEAL = Betula alleghaniensis, CACA = Carpinus caroliniana, CAOV = Carya ovata, FAGR = Fagus grandifolia, FRPE = Fraxinus pennsylvanica, LITU = Liriodendron tulipifera, NYSY = Nyssa sylvatica, PIAB = Picea abies, PIST = Pinus strobus, PRSE = Prunus serotina, QURU = Quercus rubra, SAAL = Sassafras albidum, # Pts = Number of grid points of occurrence, # Trees = Number of trees, Tr/ha = Number of trees per hectare, BA = basal area, Rel Dom = Relative Dominance, Rel Dens = Relative Density, Rel Freq = Relative Frequency, IV = Importance Value (Sum of Rel Dom, Rel Dens and Rel Freq). # BA BA Rel Rel Rel Spp # Pts Trees Tr/ha (m2) (m2/ha) Freq Dens Dom IV ACRU 19 28 96.74 6.33 21.87 27.54 30.43 13.35 71.32 LITU 12 13 44.92 5.10 17.62 17.39 14.13 23.19 54.71 PRSE 9 11 38.02 1.01 3.49 13.04 11.96 5.42 30.42 QURU 6 9 31.09 1.79 6.18 8.70 9.78 11.78 30.26 NYSY 6 9 31.09 0.43 1.49 8.70 9.78 2.80 21.28 FAGR 5 5 17.26 0.64 2.21 7.25 5.43 7.53 20.21 PIAB 2 3 10.36 0.57 1.97 2.90 3.26 11.18 17.34 SAAL 3 5 17.26 0.32 1.11 4.35 5.43 3.78 13.56 FRPE 1 1 3.47 0.16 0.55 1.45 1.09 9.75 12.29 BEAL 2 3 10.36 0.22 0.76 2.90 3.26 4.30 10.46 PIST 1 2 6.90 0.13 0.45 1.45 2.17 3.94 7.56 ACSA 1 1 3.47 0.02 0.07 1.45 1.09 1.47 4.01 CAOV 1 1 3.47 0.01 0.035 1.45 1.09 0.83 3.37 CACA 1 1 3.47 0.01 0.035 1.45 1.09 0.67 3.21 Totals 68 91 317.89 16.73 57.84 100.02 99.99 99.99 300.00 decreased from seedling through the 5 cm size class, followed by an increase in the 7 cm size class and an absence of individuals in the 9 cm size class. A similar trend was seen in Fraxinus pennsylvanica (Figure 2G), and Liriodendron tulipifera (Figure 2H). Pinus strobus densities (Figure 21) never exceeded 100 stems. ha-1 in any size class, and had very low numbers of larger trees. Picea abies (Figure 2J) included seedlings, small saplings and larger trees but no larger saplings or smaller trees. Quercus muehlenbergii (chinkapin oak) (Figure 2K) was present only as seedlings and smaller saplings. Finally, Sassafras albidum (sassafras) (Figure 2L), an understory tree that occasionally reaches the canopy, generally showed a decrease in stems numbers from seedlings to smaller trees. DISCUSSION According to Herman et al. (2001), communities with FQI values above 50 are very rare and "... represent a significant component of Michigan native bio

Page  125 ï~~2010 THE MICHIGAN BOTANIST 125 TABLE 4. Tree species composition for upland sampling points in woods at Tower Hill Camp and Retreat Center, Sawyer. Michigan. ACRU = Acer rubrum, ACSA = Acer saccharum, FAGR = Fagus grandifolia, LITU = Liriodendron tulipifera, NYSY = Nyssa sylvatica, OSVI = Ostrya virginiana, PIST = Pinus strobus, PRSE = Prunus serotina, QUAL = Quercus alba, QURU = Quercus rubra, SAAL = Sassafras albidum, # Pts = Number of grid points of occurrence, # Trees = Number of trees, Tr/ha = Number of trees/ha, BA = basal area, Rel Dom = Relative Dominance, Rel Dens = Relative Density, Rel Freq = Relative Frequency, IV = Importance Value (Sum of Rel Dom, Rel Dens and Rel Freq). # BA BA Rel Rel Rel Spp # Pts Trees Tr/ha (m2) (m2/ha) Freq Dens Dom IV PIST 33 54 122.80 5.30 12.05 30.56 34.39 30.79 95.74 QURU 29 45 102.34 5.21 11.85 26.85 28.66 30.23 85.74 FAGR 8 13 29.57 1.75 3.98 7.41 8.28 10.15 25.84 LITU 6 7 15.93 1.90 4.32 5.56 4.46 11.06 21.08 PRSE 8 11 25.03 1.26 2.88 7.41 7.01 7.34 21.76 ACRU 9 9 20.46 0.81 1.84 8.33 5.73 4.71 18.77 ACSA 7 9 20.46 0.66 1.50 6.48 5.73 3.84 16.05 SAAL 3 3 6.82 0.08 0.18 2.78 1.91 0.44 5.13 NYSY 2 3 6.82 0.09 0.20 1.85 1.91 0.54 4.30 OSVI 2 2 4.53 0.05 0.11 1.85 1.27 0.27 3.39 QUAL 1 1 2.29 0.11 0.25 0.93 0.64 0.64 2.21 Totals 108 157 357.05 17.22 39.16 100.01 99.99 100.01 300.01 diversity and natural landscapes." With an FQI > 50, the woods at THC has a sizeable number of plant species that were present in similar habitats prior to settlement by non-native people and is worthy of preservation. Given that most of the trees aged from increment cores were about 70 to 80 years old makes this a young forest compared to that at nearby Warren Woods State Park where some trees exceed 300 years old (Poulson and Platt 1996). Thus the forest at THC is a second growth forest, one that probably had been harvested twice, once in the mid 1860s at the time of the first major logging in southwestern Michigan (Ellis 1880; Coolidge 1906) and then again around 1920. These woods, located in the southwest corner of Michigan, are within the southern mesic hardwood forest type (Kost et al. 2007). As such they should have Fagus grandifolia and Acer saccharum as codominant tree species. Some nearby forests which have been studied show this pattern of dominance (Cain 1935; Gysel 1951; Woods 1979; Poulson and Platt 1996). In the THC forest, neither Fagus grandifolia nor Acer saccharum is a dominant species. Instead this role has been taken over by Quercus rubra, Pinus strobus and Acer rubrum. This compares somewhat favorably to some of the forests at nearby Warren Dunes State Park. All eight stands sampled there had Quercus rubra as a dominant, however only two stands had any Acer rubrum (Smith and Woodland 2007), probably because they were in locations more arid than those at THC.

Page  126 ï~~126 THE MICHIGAN BOTANIST Vol. 49 126 THE MICHIGAN BOTANIST Vol. 49 A. Acer rubrum B.F grandifolia Siz Clas i oits(d h r C. Acer s chorum D. Pro r so fna Size Class dpo ts (db in cm Size ClassMid ointkphb, cry E. Nyss ylvtc Quercus rubra [7,Z S 1.CI ns( tj m) Size Cs Midpoint h n m) FIGURE 2. Stem density of tree species at Tower Hill Camp and Retreat Center, Berrien County, Sawyer, Michigan. S = Seedlings (< 30 cm tall), 1 = Saplings > 30 cm tall and < 2 cm dbh, 3 = Saplings 2.00 to 3.99 cm dbh, 5 = Saplings 4.00 to 5.99 cm dbh, 7 = Saplings 6.00 to 7.99 cm dbh, 9 = Saplings 8.00 to 9.99 cm dbh, 17.5 = Trees 10.00 to 24.99 cm dbh and > 25 = Trees 25+ cm dbh. Tree density (341.58 trees ha-1) at THC was lower than that for Warren Dunes State Park (432.9 to 781.0 trees - ha-1) reported by Smith and Woodland (2007). This is to be expected since Smith and Woodland (2007) included stems as small as 8 cm dbh, while I included stems no smaller than 10 cm dbh. When the density of 8-10 cm saplings (- 120. ha-'r) from THC is added to the tree density, the total density (-461 trees - ha-1) falls at the lower end of the

Page  127 ï~~2010 THE MICHIGAN BOTANIST 127 2010 THE MICHIGAN BOTANIST 127 Fraxinspenynic H. Lrio ndron tlipfer E Size ClamssMidpoint (db in cm) 1. Pints strobus Size Class idoi t h ir cm) SPIcea be...,.... ' fi y 1. 1. Size s idpont (dbh in n) K ercusn len erli Size Class Midpints (db nc LSassafras lbidum v.,::, Ma. kM W. vn. h!n '. eC;. '.. _ 4 Y Sit Classidpoint(dbhincm S5 7I 9 Size Class Midpoints (dbh in cm FIGURE 2. Continued. range at Warren Dunes State Park. In addition, Smith and Woodland (2007) sampled an area that contained a greater range of tree ages, a greater diversity of species and a greater range of topographic conditions (e.g., elevational, slope and aspect differences), all of which may have influenced the density of trees. Acer rubrum is a species indicative of wet areas (Barnes and Wagner 2004). At THC it reached its peak abundance in places where there was standing water

Page  128 ï~~128 THE MICHIGAN BOTANIST Vol. 49 in Spring, 2009. Its shade tolerance (Barnes and Wagner 2004), its high fecundity (over 2000 seedlings * ha-1) and size class structure (Figure 2A) support the hypothesis that this species will be a canopy dominant on mesic sites for some time into the future under the current climatic conditions; however, it is considered by some to be a sublclimax species, and would be replaced by other species should the area become more arid (Russell and Yawney 1990). Nevertheless, it is shade tolerant and an aggressive colonizer of disturbed sites (Russell and Yawney 1990) and may expand further into the upland sites after disturbances that create gaps in the forest. The dominance of Quercus rubra seems to be a characteristic of forests adjacent to Lake Michigan that is not shared by forests more inland (Smith and Woodland 2007), such as the forest at Warren Woods State Park (Cain 1935; Gysel 1951; Woods 1979; Poulson and Platt 1996). Because the forest at THC is young, it is inhabited by early to middle stage successional species, Quercus rubra being one such species (Barnes and Wagner 2004). Quercus rubra is also well adapted to the mesic conditions present at THC. Its size class structure (Figure 2F), which shows a general decline in density from seedling to tree size classes, indicates its persistence as a canopy dominant. The high density of 7-cm dbh individuals preceded by the near absence of 9-cm dbh stems may have been due to an absence of a mast crop (acorns) in the several earlier years represented by the 9-cm dbh size class. This irregular production of a mast crop is a characteristic of Quercus rubra (Sander 1990) resulting from the natural selection for phenotypes that produce an unpredictable food supply for predators. It also may be due to some factor that increased mortality of the mast crop, seedlings or saplings during the years represented by the 9-cm dbh size class. Fagus grandifolia existed at low frequency and had a distribution restricted primarily to the shallow slopes at the southwestern edge of the forest at THC. Because the larger of these trees were about 125 years old, they had been spared when the forest was last logged about 80 years ago. The size class distribution of Fagus grandifolia (Figure 2B) shows a species that is reproducing itself and may persist over time, however, it may not spread east into the adjacent lower and wetter area because its seeds germinate poorly on wet sites and because it is less tolerant of flooding than Acer rubrum (Tubbs and Houston 1990). It could, however, become a canopy dominant in the upland area to the north and east of its current location. While Acer saccharum also existed at low frequency, it was scattered throughout the forest. While its size class distribution (Figure 2C) shows a decrease in density with increase in size, implying continued reproduction and future replacement, it will be a long time before it replaces the less shade tolerant species. Pinus strobus had the highest Importance Value of all species on upland sites within the forest at THC (Table 2). While its seedling and sapling densities were lower than those for Quercus rubra and Acer rubrum, it is reproducing itself, but only in openings since its seedling survival is low in shade (Wendel and Smith 1990). Even though I intentionally excluded from my samples those Pinus strobus trees that were in obvious rows, some of the canopy pines

Page  129 ï~~2010 THE MICHIGAN BOTANIST 129 that I did sample may have been among those that were planted about 70-80 years ago in the northern portion of the study area, after that had been clear cut and/or burned. Because Prunus serotina is a shade intolerant species (Marquis 1990), only those saplings found in a gap created by the demise of a canopy tree will be released from growth inhibition. In gaps, their rapid growth may get them to be canopy trees, maintaining this species at a low density. Picea abies was undoubtedly introduced as a local planting on the eastern edge of the woods, the larger trees being those that were originally planted. Based on the age of three trees that were cored, these trees were planted about 80-90 years ago. The absence of stems in all but the smallest size classes (Figure 2J) may be due to an absence of reproduction as the planted trees matured to reproductive age. Another example of this pattern existed for Liriodendron tulipifera (Figure 2H) in which the three larger sapling classes (5 cm, 7 cm and 9 cm) are empty. Finally, the abundance of Fraxinus pennsylvanica seedlings and small saplings and the paucity of individuals in the tree size classes should be noted (Figure 2G). Few seed source Fraxinus pennsylvanica trees were found within the study area, so seeding most probably came from trees planted in the adjacent residential neighborhoods. Present in the canopy were species of trees listed as shade intolerant to mid tolerant by Barnes and Wagner (2004). These include Prunus serotina, Quercus rubra, Liriodendron tulipifera and Pinus strobus. Their presence among the canopy species and their existence in the smaller size classes is most probably a feature of younger forests. As this forest matures, these species will likely be replaced by shade tolerant ones such as Fagus grandifolia and Acer saccharum. This forest then will be more like the climax forests that were here pre-settlement and more like the forest at Warren Woods State Park. Quercus muhlenbergii and Sassafras albidum are two understory species whose size class structure (Figure 2K, L) shows the effects of canopy closure as the dominant tree species mature. Quercus muhlenbergii, which is shade intolerant (Barnes and Wagner 2004), may persist as saplings, but will never become important in the forest. That this species occurs within the woods at THC may be due to recent introductions. As for Sassafras albidum, its presence in the understory seems assured since it is shade tolerant as a seedling and smaller sapling (Barns and Wagner 2004). Its presence among the canopy trees will probably diminish as the canopy trees overtop the few of these taller sassafras individuals. Judging from the size class structure of the trees in this forest which indicates that many of the tree species are reproducing themselves, this high quality (as determined from the FQI) stand appears to be in good health and will persist for some time into the future. ACKNOWLDGEMENTS I thank Robert Beemer of Deer Creek Open Spaces Association for making me aware of the need for this study, Greg Briggs, Interim Site Manager at Tower Hill Camp and Retreat Center, for per

Page  130 ï~~130 THE MICHIGAN BOTANIST Vol. 49 mission to use the woods at the camp as a place to do research and an anonymous reviewer for suggestions that made this paper more lucid. LITERATURE CITED Barnes, B.V., and W.H. Wagner, Jr. (2004). Michigan trees: a guide to the trees of the Great Lakes region. University of Michigan Press, Ann Arbor, MI. 447 pp. Billington, C. (1924). The flowering plants and ferns of Warren Woods, Berrien County, Michigan. Papers of the Michigan Academy of Science, Arts and Letters 4Z: 81-109. Brewer, L.G., T.W. Hodler and H.A. Raup. (1984). Presettlement vegetation of southwestern Michigan. Michigan Botanist 23: 153-156. Brewer, R. (1980). A half-century of changes in the herb layer of a climax deciduous forest in Michigan. Journal of Ecology 68: 823-832. Brewer, R., and P.G. Merritt. (1978). Wind throw and tree replacement in a climax beech-maple forest. Oikos 30: 149-152. Cain, S.A. (1935). Studies on virgin hardwood forest: III. Warren's Woods, a beech-maple climax forest in Berrien County, Michigan. Ecology 16: 500-513. Coolidge, O.W. (1906). A twentieth century history of Berrien County Michigan. Lewis Publishing Company, Chicago, IL. 1007 pp. Curtis, J.T. (1959). The vegetation of Wisconsin: an ordination of plant communities. The University of Wisconsin Press, Madison, WI. 657 pp. Dickmann, D.I., and L.A. Leefers. (2003). The Forests of Michigan. The University of Michigan Press, Ann Arbor, MI. 397 pp. Donnelley, G.T., and P.G. Murphy. (1987). Warren Woods as forest primeval: A comparison of forest composition with pre-settlement beech-sugar maple forests of Berrien County, Michigan. Michigan Botanist 26(1): 17-24. Ellis, F. (1880). History of Berrien and Van Buren Counties Michigan. D.W. Ensign & Co., Philadelphia, PA. 548 pp. Gleason, H.A., and A. Cronquist. (1991). Manual of the Vascular Plants of the United States and Adjacent Canada, 2nd ed. New York Botanical Garden, Bronx, New York. 910 pp. Greenwood, J.J.D. (1996). Basic techniques, pages 11-110. In W. J. Sutherland, Ecological Census Techniques. Cambridge University Press, New York, NY. 336 pp. Gysel, L.W. (1951). Borders and openings of beech-maple woodlands in southern Michigan. Journal of Forestry 49: 13-19. Herman, K.D., L.A. Masters, M.R. Penskar, A.A. Reznicek, G.S. Wilhelm, W.W. Brodovich, and 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, MI. 19 pp. + Appendices. Kost, M.A., D.A. Albert, J.G. Cohen, B.S. Slaughter, R.K. Schillo, C.R. Weber and K.A. Chapman. (2007). Natural communities of Michigan. Michigan Natural Features Inventory, Report Number 2007-21, Lansing, Michigan. 314 pp. Larson, J.D. (1980). Soil survey of Berrien County, Michigan. USDA Soil Conservation Service. Washington, DC. 192 pp. Marquis, D.A. (1990). Prunus serotina Ehrh. In: Burns, R.M., and B.H. Honkala, technical coordinators. Silvics of North America, Volume 2, Hardwoods. USDA Forest Service Agriculture Handbook 654. Washington, DC. p. 560-569. Poulson, T.L., and W.J. Platt. (1996). Replacement patterns of beech and sugar maple in Warren Woods, Michigan. Ecology 77: 1234-1253. Russell, W.S., and H.W. Yawney. (1990). Acer rubrum L. In: Burns, R.M., and B.H. Honkala, technical coordinators. Silvics of North America, Volume 2, Hardwoods. USDA Forest Service Agriculture Handbook 654. Washington, DC. p. 60-69. Sander, I.A. (1990). Quercus rubra L. In: Burns, R.M., and B.H. Honkala, technical coordinators. Silvics of North America, Volume 2, Hardwoods. USDA Forest Service Agriculture Handbook 654. Washington, DC. p. 727-733. Smith, P.F., and D.W. Woodland. (2006). Vascular plant study of Warren Dunes State Park, Berrien County, Michigan. Michigan Botanist 45(1): 1-58.

Page  131 ï~~2010 THE MICHIGAN BOTANIST 131 Smith, P.F., and D.W. Woodland. (2007). Forest composition study of the Great Lakes coastal forest at Warren Dunes State Park, Berrien County, Michigan. Michigan Botanist 46(2): 33-62. Tubbs, C.H., and D.R. Houston. (1990). Fagus grandifolia Ehrh. In: Burns, R.M., and B.H. Honkala, technical coordinators. Silvics of North America, Volume 2, Hardwoods. USDA Forest Service Agriculture Handbook 654. Washington, DC. p. 325-332. Wells, J.R., and P.W. Thompson. (1982). Plant communities of the sand dunes region of Berrien County. Michigan Botanist 21(1): 3-38. Wendel, G.W., and H.C. Smith. (1990). Pinus strobus L. In: Burns, R.M., and B.H. Honkala, technical coordinators. Silvics of North America, Volume 2, Hardwoods. USDA Forest Service Agriculture Handbook 654. Washington, DC. p. 476-488. Woods, K.D. (1979). Reciprocal replacement and the maintenance of codominance in a beech-maple forest. Oikos 33: 31-39.