Kevin A. Yokum, Ted R. Angradi, and Donald C. Tarter.
Ecology of Peltoperla arcuata and Tallaperla maria (Plecoptera: Peltoperlidae) at the Fernow Experimental Forest, Tucker County, West Virginia.
Psyche 102:151-168, 1995.

Full text (searchable PDF, 992K)
Durable link: http://psyche.entclub.org/102/102-151.html

The following unprocessed text is extracted from the PDF file, and is likely to be both incomplete and full of errors. Please consult the PDF file for the complete article.

ECOLOGY OF PELTOPERLA ARCUATA AND TALLAPERLA MARIA (PLECOPTERA: PELTOPERLIDAE) AT THE FERNOW EXPERIMENTAL FOREST, TUCKER COUNTY,
WEST VIRGINIA
We examined the abundance, life history, and production of the stoneflies Peltoperla arcuata and Tallaperla maria (Plecoptera: Peltoperlidae) in four forested headwater streams at the Fernow Experimental Forest, Tucker County, West Virginia. Peltoperla arcuata was most abundant in the smallest watersheds (<lo0 ha), and was present at all sites. Tallaperla maria was most abundant in watersheds >200 hectares (ha), was restricted to sites with a base- flow alkalinity of >2 mg L-I CaC03, and was the dominant peltop- erlid only at sites with an alkalinity >15 mg L-l. We conclude that water chemistry overrides stream size as a determinant of species- specific distribution of Fernow peltoperlids. Both taxa had semi- voltine life cycles with an 18-month naiadal period following a 6-month egg diapause. Emergence was during May-July for both species. Peltoperla arcuata had about 15 instars; T. maria had about 14 instars. Peltoperlid production was highest (509 mg m-2 y-1) in a 128 ha watershed where only P. arcuata was collected; P. arcuata production was lowest (17 mg m-= y I) in a 4th order stream (1536 ha). Tallaperla maria production was highest (271 m-2 y-1) in a 257 ha watershed partially underlain by limestone. Production across streams was higher for P. arcuata (205 mg m-2 y-l) than for T maria (91 mg m-2 y-I).
Stoneflies (Plecoptera) of the family Peltoperlidae are "shred- der-detritivores" (Merritt and Cummins, 1996) which comminute '~e~artment of Biological Sciences, Marshall University, Huntington, West ginia 25755
2~orresponding author, USDA Forest Service, Northeastern Forest Experiment Sta- tion, Parsons, West Virginia 26287
Manuscript received 9 October 1995.




================================================================================

152 Psyche [vo~. 102
leaves and other coarse allochthonous organic matter into finer par- ticles that can be ingested by other macroinvertebrates in headwa- ter streams. Shredders are therefore an important trophic link between terrestrial and aquatic components of forest ecosystems (Cummins, Wilzbach, Gates, Perry, and Taliaferro, 1989). Shredder assemblages in certain Appalachian headwater habitats, especially leaf packs in riffles, are often dominated by peltoperlids (Griffith, Perry, and Perry, 1994).
Peltoperlidae is one of the smallest families of Plecoptera, com- prised of 14 genera and about 40 species (Stark and Stewart, 1981). Of the six Nearctic genera, three are eastern: Peltoperla, Tallaperla, and Viehoperla (Stewart and Stark, 1993). Naiads of Peltoperla and Tallaperla have identical gill formulae, and similar habitus and mouthparts. No species-specific characters were found for Peltoperla or Tallaperla naiads in early instars by Stark and Kondratieff (1987). The characteristic used to distinguish between naiads of the two genera is the presence of black spots on the pronota and mesonota of P. arcuata. Eastern peltoperlids have a very distinctive "roach-like" habitus. We speculate that this body shape is an adaptation for penetration between tightly appressed leaves comprising the leaf packs within which peltoperlids are often very abundant.
Both taxa are widespread in Appalachia (Stewart and Stark, 1993). At the Fernow Experimental Forest, Tucker County, West Virginia, P. arcuata and T. maria occur together at some sites; at other sites only one species is found. The basis for habitat segrega- tion between these species is unknown.
The objectives of this study were to document the life history and secondary productivity of P. arcuata and T. maria on the Fer- now Experimental Forest, and to examine habitat factors poten- tially underlying species-specific distribution patterns. The Fernow Experimental Forest (USDA Forest Service) is located in Tucker County, West Virginia (39'03'N, 79'40' W) . The 1900 hectare (ha) forest includes the entire watershed of Elklick Run, a fourth order tributary of the Black Fork of the Cheat River (Fig. 1). Located in the unglaciated Allegheny Plateau Province of the Central Appalachians, the Fernow Experimental Forest receives



================================================================================

Yokurn, Angradi & Tarter
Fig. 1. Map of Fernow Experimental Forest, Tucker Co., WV. Numbers refer to study sites as described in Table 1.




================================================================================

154 Psyche [vo~. 102
about 150 cm of annual precipitation. Elevation ranges from 554 m to 1113 m.
The Fernow is primarily mixed deciduous second-growth forest. Important tree species include northern red oak (Quercus rubra L.), sugar maple (Acer saccharum Marsh.), yellow-poplar (Lirio- dendron tulipifera L.), American beech (Fagus grandifolia Ehrh.), black cherry (Prunus serotina Ehrh.), American basswood (Tilia americana L.), and white ash (Fraxinus americana L.). The western half of the Fernow (roughly west of Elklick Run) is underlain by shales and sandstones of the Hampshire formation and the eastern half is underlain by limestones of the Greenbrier formation. Adams, Kochenderfer, Wood, Angradi, and Edwards (1994) present additional information about the Fernow Experi- mental Forest.
Four streams were sampled in this study: Elklick Run, Camp Hollow, Wilson Hollow, and Canoe Run (Fig. 1). Stream substra- tum is primarily flat, subangular sandstone cobble and gravel. Bedrock outcrops are common, especially in Elklick Run; large woody debris is abundant, especially in streams draining smaller watersheds.
Peltoperlid naiads were collected from October 1993 to Septem- ber 1994 using a Surber sampler (0.10 m2, 0.25-mm mesh). Across the four streams, 23 sites were sampled approximately monthly with two samples taken at each site (Table 1, Fig. 1). Samples were preserved in 5% formalin plus a small amount of phloxine-B stain. Naiads were generally found in leaf packs in riffles; in summer when leaf packs were absent, naiads were found deeper in riffle substrates.
At each sample site during monthly collections, stream width and sample depth and velocity were measured. Electronic thermo- graphs continuously recorded temperatures in each stream. Monthly water samples collected from selected sites on each stream were analyzed for pH, conductivity, alkalinity, calcium, and nitrate.
Naiads were measured using a computer digitizing system and measuring software (Java, Jandel Scientific). Naiads were placed under a video camera where their image was captured on a frame



================================================================================

19951 Yokum, Angradi & Tarter 155
freezing monitor, and head width (widest portion posterior to eye) and total length (cerci excluded) were measured. Additional details are presented in Yokum (1995).
The number of instars for each species was estimated by the Janetschek (1967) method in which a sliding mean of three size classes was used to resolve positive peaks in the head-width fre- quency distribution. Positive peaks are putative instars. Stewart and Stark (1993) consider this method useful despite its shortcomings. Selected specimens of known head width were dried (60å¡C 48 h) and weighed to the nearest 0.001 mg. Dry mass of the remaining naiads was estimated using a least-squares regression of log-transformed dry mass on log-transformed head width. Because slopes of regressions were not different among streams, equations based on collections from all streams were used (Table 2). Sec- ondary productivity was estimated for P. arcuata and T. maria in each stream (all sites combined) using the size-frequency method corrected for the cohort production interval (Benke, 1996). RESULTS AND DISCUSSION
Habitat and water quality
For each stream, stream width, sample depth, and sample-point velocity, generally increased with increasing watershed area (Table 1). Among streams Elklick Run represents the largest stream sam- pled; Wilson Hollow was the smallest stream sampled. Physical habitat conditions at Camp Hollow and Canoe Run were similar. Among-stream differences in pH, conductivity, alkalinity, and calcium concentration reflect watershed geology. Sites at Elklick Run and lower Canoe Run partially drain limestone geologic for- mations, and had pH values >7.0, alkalinities >15 mg L-1, and cal- cium concentrations >7 mg L-l (Table l). Other sites had pH values ~7.0, and alkalinities < 7 mg L1. Nitrate concentrations and total degree days were similar among streams (Table 1). Weekly mean temperatures were similar among streams and sites (Fig. 2). Abundance and Distribution
Peltoperlid abundance varied greatly among streams and among sites (Fig. 3). Mean annual abundance of peltoperlids was highest at Wilson Hollow (180 m-2), where only P. arcuata was collected. There was a trend towards increasing abundance at the more



================================================================================

Table 1. Physical habitat and water quality characteristics of study sites on the Fernow Experimental Forest. Values are means for the study period based on monthly samples. Water quality data were not collected at all sites. Sample Sample Stream Alkalinity Annual
WS area velocity depth width Conductivity (mg L-I) Calcium Nitrate degree Stream Site (ha) (m s-I) (cm) (m) pH (pS cm-I) CaCOi (mg L-I) (mg L-I) days Wilson Hollow 1
2
3
4
5
6
Camp Hollow 1
2
3
4
5
6
-
a Stream drains experimentally acidified watershed.



================================================================================

Table 1. Continued
- -
Sample Sample Stream Alkalinity Annual
WS area velocity depth width Conductivity (mg L") Calcium Nitrate degree Stream Site (ha) (m s-l) (cm) (m) pH (pS cm-I) CaCO, (mg L-I) (mg L-') days Canoe Run 1 257 1.4
2 239 1.2
3 203 1.1
4 51 1.0
5 26 0.9
Elklick Run 1 1536 1.6
2 1407 1.4
3 1387 1.4
4 818 1.4
5 313 1.1
6 99 1.1
a Stream drains experimentally acidified watershed.



================================================================================

Psyche
Table 2. Regression equations for the relationships between log-transformed head width (HW, mm) and log-transformed dry weight (W, mg) for P. arcuata and T maria on the Fernow Experimental Forest.
Species n Regression r2
P. arcuata 79 Ln(W) = -0.72+3.33 Ln(HW) 0.94 T maria 53 Ln(W = -0.77+3.09 Ln(HW) 0.96 D J F M A M J J A S O N D J
1 994
Fig. 2. Weekly mean temperature for Fernow study sites. Numbers in legend refer to study sites in Fig. 1 and Table 1.




================================================================================

Yokum, Angradi & Tarter 159
T. maria
\
500
::;
0
500
400 - Canoe Run \
300 -
200 -
100 -
0,
,"-.--, IÌÔIÌ
400,-
1 2 3 4 5 6
Sites
250
Elklick Run n I? mcuata
Elklick
Canoe
Camp Wilson
Run
Run
Hollow Hollow
Fig. 3. Mean density of peltoperlid naiads at Fernow study sites and for sites at each stream combined. Sites are as in Fig. 1 and Table 1. There was no site 6 at Canoe Run.




================================================================================

160 Psyche [vo~. 102
upstream sites. Mean annual abundance of peltoperlids was lowest at Elklick Run (38 m-2), of which most (32 m-2) were T. maria. At Elklick Run, Peltoperla arcuata was relatively more abundant at the most upstream sites. At Camp Hollow there was also an upstream trend in abundance of P. arcuata. Tallaperla maria was common only at the two downstream sites at Camp Hollow. Tal- laperla maria was abundant at the three downstream sites at Canoe Run. Peltoperla arcuata was rare but present at all Canoe Run sites, exclusively so at the most upstream sites (Fig. 3). Because the largest streams, Elklick Run and Canoe Run, also had the highest baseflow alkalinities (Table I), it is difficult to determine the relative importance of stream size and water chem- istry in the distribution of peltoperlids in Fernow streams. At Elk- lick Run, T. maria and P. arcuata were both present with T. maria dominant at all sites despite great variation in stream size but with uniformly high alkalinity. At Canoe Run, T. maria was present only at the downstream, more alkaline sites. Peltoperla arcuata was present at the most upstream sites at Canoe Run, but did not thrive there as it did in similar sized sites at Wilson Hollow a few km away.
Across sites for which we have water quality data, the percent of total peltoperlid abundance comprised of T. maria (relative abundance) was strongly correlated with alkalinity (Fig. 4). Rela- tive abundance of T. maria was also correlated with watershed area with two exceptions: at two similar-sized streams, lower Canoe Run and lower Camp Hollow, relative abundance of T. maria was greater at Canoe Run, the more alkaline stream (Figs. 3 and 4, Table 1); and T. maria was the dominant peltoperlid at the smallest Elklick Run site (99 ha); non-alkaline streams of a similar size are dominated by P. arcuata elsewhere on the Fernow. We conclude that on streams of the Fernow Experimental For- est, P. arcuata thrives and is the dominant peltoperlid only in small streams (e.g., ~200 ha) with low base flow alkalinity (e.g., c 2 mg L-I), and T. maria is the dominant peltoperlid in more alkaline streams (>I5 mg L-I). At intermediate sites such as Camp Hollow sites 1 and 2 (182-199 ha, 6 mg L-I alkalinity), both species occurred but neither was abundant (Figs. 3 and 4). Fernow streams with a higher alkalinity (>30 mg L-I) are overwhelmingly domi-



================================================================================

19951 Yokum, Angradi & Tarter 161
0 5 10 15 20 25 30 0 600 1200 1800
Alkalinity (mg1L) Watershed area (ha)
Fig. 4. Pearson correlations between relative abundance of T. maria (T. maria abundance + total peltoperlid abundance * 100) and baseflow alkalinity and between relative abundance of T. maria and watershed area. Symbols denote stream: (0) Wilson Hollow, (0) Camp Hollow, (V) Canoe Run, (0) Elklick Run. Some symbols hidden; n=13.
nated by the omnivorous amphipod Gammarus minus, and peltop- erlids are uncommon (T. Angradi, unpublished data). Miller and Kovalak (1979) speculated that density of P. arcuata increased in the upstream direction in a small Pennsylvania stream partly because smaller streams (with their smaller watersheds) were less subject to high flows which could scour naiads and the leaves on which they feed. They reasoned that although total organic matter was roughly equivalent in all channels on an annual basis, organic matter was retained longer in the smaller channels which are less subject to high spring flows. We accept this explana- tion for the increased abundance of P. arcuata with decreasing watershed area at Wilson Hollow and Camp Hollow, and we add that I? arcuata probably has adaptations for burrowing into the deep interstitial zone, because these upper stream reaches are usu- ally without surface flow in late summer and autumn. Life history
Peltoperla arcuata and T. maria exhibited semivoltine life cycles (Fig. 5). Peltoperla arcuata, for which we have the most data, had an approximately 18 month larval period following an egg diapause of about 6 months. Late instars of T. maria were rare in our collections, but we infer a life cycle similar to P. arcuata.



================================================================================

Psyche [vo~. 102
Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sap T. maria
Oct Nov Oec Jan Feb Mar Apr May Jun Jui Aug Sap 1993 1 994
Fig. 5. Semivoltine life cycle of peltoperlids for all Fernow sites. Numbers adja- cent to horizontal axes are sample sizes. E=emergence.



================================================================================

19951 Yokum, Angradi & Tarter 163
The mesh size of the net we used (0.25 mm; the smallest naiads we collected had head widths <0.2 mm) may have failed to adequately sample first instars which probably appear in November-January (O'Hop, Wallace, and Haefner, 1984; Elwood and Cushman, 1975). Overlapping cohorts of T. maria naiads were collected from March through April (Fig. 5); overlapping cohorts P. arcuata naiads were collected from February through July (Fig. 5). Instar analysis suggested that P. arcuata had at least 15 instar peaks based on head width (Fig. 6), and T. maria had about 14 instar peaks. Final instar size was larger for P. arcuata. However, Head width (mm)
Fig. 6. Frequency distribution of peltoperlid naiads for all Fernow sites and (lower panels) instar analysis in which the actual number of naiads in each size class was subtracted from the sliding mean of the number of naiads in three size classes (mean of the size class in question and the next smallest and next largest size class). Positive peaks are putative instars.



================================================================================

164 Psyche [vo~. 102
as noted above, we collected few late instar T. maria naiads. Rug- gles and Tarter (1991) estimated 10-12 instars for P. tarteri. Emergence of P. arcuata occurred from late May through July based on the loss of the larger cohort naiads after April (Fig. 5). Emergence trap data collected in 199 1 supports this conclusion (Michael Griffith, unpublished data). We did not collect mature T. maria naiads after April, but emergence probably occurred in May- June. In a Maryland stream at a latitude similar to the Fernow Experimental Forest, Duffield and Nelson (1990) reported emer- gence of T. elisa in May and June. Elwood and Cushman (1975) reported emergence of T. maria in April and May in a Tennessee stream. T. maria emerge in May at Coweeta in North Carolina (O'Hop et al., 1984).
Peltoperlids apparently exhibit indeterminant voltinism. Semi- voltinism of T. maria in North Carolina was reported by O'Hop et al. (1984) and Huryn and Wallace (1987), and in Tennessee by Elwood and Cushman (1975). Miller and Kovalak (1979) reported a univoltine life cycle for P. arcuata in Pennsylvania. Claassen (193 1) reported a semivoltine life cycle in New York. In contrast to the semivoltine peltoperlids of the Fernow Experimental Forest, Ruggles and Tarter (1991) reported a univoltine life cycle for I? tarteri at Paint Creek in Fayette County, West Virginia. Emergence of P. tarteri occurred in June and naiads hatched in July. The naiads grew slowly until October and then rapidly through fall and winter. Univoltinism of P. tarteri may be related to higher stream temperatures or different food availability at the lower elevation stream examined by Ruggles and Tarter (1 991). At the Fernow Experimental Forest, stream flow is lowest in July-October. Surface flow ceases in many stream reaches in autumn. Also, leaf litter is scarce in perennial reaches by late sum- mer (T. Angradi, unpublished data). Semivoltine life cycles with an egg diapause may allow P. arcuata and T. maria naiads to avoid these unfavorable environmental conditions in late summer and early autumn.
Production
Production was calculated using a cohort production interval (CPI) of 550 days because life cycles of both taxa were semivol- tine (18 month larval period) (Table 3). For T. maria, production



================================================================================

19951 Yokum, Angradi & Tarter 165
Table 3. Secondary production estimates for Peltoperlidae at the Fernow Experi- mental Forest and at the Coweeta Hydrologic Laboratory, near Franklin, NC. CPI = 550 (Fernow) or 540 (Coweeta).
WS area Production
Stream Taxa (ha) (mg m-2y1)1 Source
Fernow Experimental Forest, WV
Elklick Run T maria
l? arcuata
Canoe Run T maria
l? arcuata
Camp Hollow T maria
P. arcuata
Wilson Hollow T. maria
P. arcuata
WS 42 P. arcuata
WS 3 P. arcuata
This study
This study
This study
This study
This study
This study
This study
This study
Griffith et al. 1994
Griffith et al. 1994
Coweeta Hydrologic Laboratory, NC
Upper Ball Creek Peltoperlidae4 39 299
Huryn and Wallace 1987
Grady Branch T maria
13 496 O'Hop et al. 1984
Sawmill Branch T maria
9 560 O'Hop et al. 1984
C55 Peltoperlidae
8 739 Lugthart and Wallace 1992
1 Production values for Fernow are dry weight basis; Coweeta values are ash-free dry weight.
2 WS 3 and WS 4 roughly correspond to Camp Hollow sites 5 and 6. Values adjusted for a CPI of 550 days.
Includes T maria, T anna, and Viehoperla ada. Mean of two years.
was highest in Canoe Run (271 mg m-2 y-l) and lowest in Camp Hollow (13 mg m-2 y-l; T. maria was not found at Wilson Hollow). For P. arcuata, production was highest at Wilson Hollow (509 mg m-2 y-I) and lowest at Elklick Run (17 mg m-2 y-I). Mean (å±SD production for all streams combined
was higher for P. arcuata
(205å±19 mg m-2 y-l) than for T. maria (91~108 mg m-2 y-1). Total