sign in sign up
<<
Home , European pond turtle, Graptemys, Riparian buffer

Journal of Environmental Management (2001) 62, 000–000 doi:10.1006/jema.2001.0454, available online at http://www.idealibrary.com on

Stream and riparian management for freshwater

J. R. Bodie

The regulation and management of ecosystems worldwide have led to irreversible loss of wildlife . Due to recent scrutiny of water policy and feasibility, there is an urgent need for fundamental research on the biotic integrity of and riparian zones. Although riverine turtles rely on stream and riparian zones to complete their life cycle, are vital producers and consumers, and are declining worldwide, they have received relatively little attention. I review the literature on the impacts of contemporary stream management on freshwater turtles. Specifically, I summarize and discuss 10 distinct practices that produce five potential biological repercussions. I then focus on the often-overlooked use of riparian zones by freshwater turtles, calculate a biologically determined riparian width, and offer recommendations for ecosystem management. Migration data were summarized on 10 species from eight US states and four countries. A encompassing the majority of freshwater migrations would need to encompass a distance of 150 m from the stream edge. Freshwater turtles primarily chose high, open, sandy to nest. Nests in North America contained eggs and hatchlings during April through September and often through the winter. In addition, freshwater turtles utilized diverse riparian habitats for feeding, nesting, and overwintering. Additional documentation of stream and riparian use by turtles is needed.  2001 Academic Press

Keywords: conservation, riparian buffer, stream management, turtles, wildlife habitat.

Introduction Recent scrutiny of water policy (e.g. Hirsch et al., 2001) and dam feasibility (e.g. World Commission on , 2000) have height- Stream management pervades not only ened interest in research and management aquatic but also terrestrial systems. For of streams, from intermittent brooks to large example, 19 300 km of United States rivers (Haeuber and Michener, 1998). Like- are actively modified for commercial navi- wise, there is a global surge of research on gation through the creation, operation and riparian ecology and management (Naiman maintenance of 75 , 276 navigation et al., 2000). Although large-scale stream locks and 13 670 km of federally maintained studies are yielding results (e.g. Molles et al., levees (US Army Corps of Engineers, 2000). 1998, Toth et al., 1998), there remains an Management of these rivers directly and urgent need for fundamental research on the indirectly affects a combined drainage area biotic integrity of streams and riparian zones 2 of 6 954 150 km or 87% of the total area (Haeuber and Michener, 1998). of the coterminous US (Statistical Abstract Much of our current understanding of of the US, 1998). Although the social ben- Division of Biological stream ecology has come from focused stud- Sciences, University of efits of these actions include hundreds of ies on fishes (e.g. Moyle and Vondracek, Missouri, Columbia, MO billions of dollars in commerce, hydropower 1985; Schlosser, 1990). Freshwater turtles 65201, USA and estimated protection from flood damages, have received comparatively little attention, Present address: Bodie environmental costs are great (Sparks et al., despite that 20 of 24 species listed under Design Group, 507 Crooked Oak Drive, 1998). Worldwide, human regulation of 60% US law are associated with fluvial habitats of total streamflow has led to irreversible loss Pawleys Island, SC 29585, (US Fish and Wildlife Service, 2000). Like- USA of species and ecosystems (World Commission wise, severe species reductions in Asia, India, on Dams, 2000). Received 23 November North America, and South America signify 1999; accepted 26 March Email of author: [email protected] a global decline of riverine turtles (Gibbons 2001

0301–4797/01/060000C000 $35.00/0  2001 Academic Press 2 J. R. Bodie

et al., 2000). In many lotic ecosystems, high Riparian use annual consumption of eggs and hatchlings by a wide range of predators (e.g. mammals, There were many anecdotal observations of , other ), and densities from 0Ð02 approximate distances of nests and nesting to 1Ð3 individuals per meter of shoreline habitats from streams, but only actual mea- indicate that freshwater turtles are vital pro- surements of individual nest and overwinter ducers and consumers (Tinkle, 1958; Plum- sites from streams were used. Data were mer, 1977; Moll, 1990; Vogt and Villareal, considered reliable if derived from direct 1993, Foscarini and Brooks, 1997). Moreover, examinations of terrestrial nest and overwin- because freshwater turtles require terrestrial ter sites adjacent to known sources of tur- habitats to complete their life cycle, they can tles. Researchers located these sites through provide more comprehensive information on direct observation, radiotelemetry, signs of riparian ecosystems. For instance, the impor- depredation (i.e. eggshells), and by following tance of semi-aquatic organisms in buffer turtle tracks originating at the water’s edge. delineation has recently been recognized for I also searched for data concerning the use isolated (e.g. Burke and Gibbons, of riparian habitats for reasons other than 1995, Semlitsch, 1998), but to my knowledge nesting and overwintering (e.g. feeding). Only has not been addressed for streams. actual measurements of seasonal migrations I review the literature on the impacts of through the use of radiotelemetry and mark- contemporary stream management on fresh- recapture methods were included. For each water turtles. Specifically, I summarize and reference, I recorded the specific timing (in discuss 10 distinct practices that produce months) and habitat association for the cor- five potential biological repercussions. I then responding riparian use. I summarized data focus on the often-overlooked use of riparian for each species on average migration dis- zones by freshwater turtles, calculate a bio- tance from the stream edge for the purposes logically determined riparian buffer width, of nesting, overwintering, or feeding. and offer recommendations for stream and riparian management. Results and discussion

Methods Management practices

I categorized literature accounts of contempo- Reliable data were obtained from published rary stream-management into ten practices literature and unpublished dissertations affecting freshwater turtles (Table 1). I dis- for studies concerning stream-management cuss aspects of each management practice practices or riparian use pertaining to fresh- individually. water turtles. Although the majority of streams worldwide are managed, the effects on freshwater turtles have been studied most frequently in the US Reduction of snags and logjams Systematic clearing of vegetative debris from many US streams began nearly two hun- Management practices dred years ago (Sedell and Froggatt, 1984). Moll (1980) reports that Illinois tur- Contemporary practices of stream-manage- tle populations suffered drastic changes in ment were summarized directly from the composition and abundance of prey species literature. Although each of these practices (e.g. invertebrates, fishes) due to clearing of is distinct, their effects may be direct or . Moll reports that woody debris indirect (e.g. human riparian uses may lead remaining along the Illinois River is impor- to or pollution). Practices were tant for basking by species such as map classified, based on the literature accounts, turtles ( spp.) and sliders (Tra- as potentially causing one or more population chemys scripta). False map turtles (Grapte- or species-level consequences. mys pseudogeographica) collected from 186 Stream and riparian management for freshwater turtles 3

Table 1. Direct and indirect stream-management practices (numbered left column) and their potential effects on freshwater turtles (lettered right column)

1. Reduction of snags and logjams—A, D, E A. Change in historical food supply 2. Riparian draining—A, B, E B. Change in population structure 3. Channelization—B, D, E C. Nest inundation and failure 4. Impoundment—A, B, D, E D. Population fragmentation 5. Flow regulation—B, C, E E. Local reduction or elimination of species 6. Reduction of sandbars and beaches—C, D, E 7. Human riparian use—A, D, E 8. Pollution and siltation—B, D, E 9. Management for monotypic conditions—B, E 10. Unsustainable use—B, D, E

The letters following each practice correspond to its biological effects in the right column. study sites in Kansas preferred habitats with River spend a substantial portion of the year basking structures (Fuselier and Edds, 1994). in flooded riparian wetlands, when these Likewise, extensive surveys of basking map habitats are available (Bodie and Semlitsch, turtles in Kentucky, Louisiana, and Missis- 2000a,b). Turtle population density as well sippi showed significant positive correlations as species diversity may be higher in flood- between turtle density and deadwood den- plain wetlands than in all other lotic habitats sity (Lindeman, 1999). Removal of snags (Timken, 1968; Tucker et al., 1999; Bodie within the Warrior River Basin in Alabama et al., 2000). Jones (1996) touted the impor- is in part responsible for fragmentation of tance of riparian cypress ponds to foraging preferred habitat and hence, populations of female yellow-blotched map turtles (Grapte- the US-listed flattened musk turtle (Ster- mys flavimaculata). Moreover, juvenile map notherus depressus; Dodd, 1990). Vandewalle turtles, sliders, and western pond turtles and Christiansen (1996) sampled seven major were more abundant in floodplain wetlands, streams and a in Iowa and found presumably taking advantage of food produc- a significant negative relationship between tivity in these shallow, fertile habitats (Reese river modification, including clearing and and Welsh, 1998a,b, Bodie and Semlitsch, snagging, and turtle species richness. Reese 2000a,b). In this regard, floodplain wetlands and Welsh (1998b) recommended the natural are nursery habitats for many turtles just as or managed addition of woody debris in the they are for many fishes (Hesse and Mestl, Trinity River in California in to provide 1993). greater habitat complexity for western pond turtles (Clemmys marmorata). In addition, the importance of vegetative debris as crit- Channelization ical to lotic systems extends beyond turtles to include their prey (e.g. fishes and inverte- Stream channelization is used worldwide brates; reviewed by Marzolf (1978), Simpson to control flooding, increase agricultural et al., 1982 and Benke et al. (1985)). acreage, improve navigability, or maintain an efficient flow of water. However, the direct and indirect effects of channelization Riparian draining include reduction of food resources, habi- tat loss, and concomitant shifts in species Draining of floodplains through construction composition (Simpson et al., 1982). Three of canal and levee systems also has a long his- species in the US, Illinois mud turtles, Bland- tory in the US and abroad (Gore and Shields, ing’s turtles, and smooth softshells ( 1995). Certain declines of Illinois mud turtles mutica), have been reduced or eliminated ( flavescens spooneri), Blanding’s from the Illinois River due to channelization turtles (Emydoidea blandingi), and map tur- (Moll, 1980). The physical process of chan- tles can be blamed on draining of riparian nel dredging may also directly disturb turtles wetlands in Illinois and Iowa (Moll, 1980, overwintering in streambed (Gra- Vandewalle and Christiansen, 1996). False ham and Graham, 1997). Channelization has map turtles and sliders along the Missouri been shown to reduce habitat diversity, which 4 J. R. Bodie

also reduces populations of western pond Flow regulation turtles that prefer a system of pools and shallows (Reese and Welsh, 1998b). Exten- One of the central purposes of stream sive sampling of freshwater turtles in Iowa impoundment is flow regulation. Stream demonstrated a significant negative relation- flows are typically regulated with respect to ship between channelization and species rich- the needs of , hydropower, navi- ness (Vandewalle and Christiansen, 1996). gation, and recreation. These needs rarely Moreover, although most riverine turtles are correspond with those of native stream flora highly mobile, some studies suggest that cer- and fauna (Kushlan and Jacobsen, 1990; tain species and sexes may avoid crossing Sparks, 1995). Similar to fishes (see Figures 9 streams with high water velocities, possi- and 10 in Galat et al., 1998), freshwater bly due to channelization (Plummer, 1977; turtles depend upon the proper coupling of Jones, 1996; Bodie and Semlitsch, 2000a,b). natural riverine hydrology and nesting habi- Therefore, channelization in some streams tat accessibility. Turtles often nest during a may isolate and fragment freshwater tur- point in the season when stream water levels tle populations, with potential demographic were historically dropping, perhaps to avoid and evolutionary effects such as reduction of nest flooding (Tucker et al., 1997). Although gene flow. turtle nests near stream shorelines doubt- less suffered mortality from inundation prior to stream regulation, contemporary stream- Impoundment flow management may place turtle nests in greater jeopardy. Tucker et al. (1997) sug- Many of the well-known effects of impound- gested that turtle embryos could suffer high ing streams on fishes and invertebrates mortality late in development when water [reviewed by Petts (1984)] apply to fresh- levels are artificially elevated for summer water turtles. The most direct consequence navigation. Conversely, in winter, juveniles of impoundment is fragmentation of habi- selecting shallow overwinter sites in the fall tat and consequently, populations. There is may suffer mortality due to artificial reduc- evidence of population insularization and tions in water level at the end of the naviga- decline for several species including ringed tion season that exposes turtles to freezing or sawback (Graptemys oculifera), flattened desiccation (Bodie and Semlitsch, 2000a,b). musk, and western pond turtles (US Fish and Wildlife Service, 1987, Dodd, 1990, Reese and Welsh, 1998a). Impoundment reservoirs also Reduction of sandbars and beaches reduce turtle diversity by homogenizing habi- tat and supporting more lentic species (Van- Sandbars and beaches have been virtu- dewalle and Christainsen, 1996; Reese and ally eliminated from many US rivers due Welsh, 1998a,b; Tucker et al., 1999). These to stream-management practices including reservoirs may also have indirect effects on channelization, impoundment, and flow reg- turtle populations including introduction of ulation. For example, islands and sandbars exotic competitors, predators and vectors for along a 740-km portion of the Missouri River disease (Vannote et al., 1980). Asian species were reduced 98% from 1879 to 1954 (Funk such as chitra and baska, and Robinson, 1974). Declines of map tur- which were already in danger due to overcol- tles and softshells in the lower Missouri lection, are now facing imminent extinction River have been attributed to loss of these due to direct and indirect effects of enor- habitats (Johnson, 1992). Like threatened mous reservoirs (van Dijk and Thirakhupt, piping plovers (Charadrius melodus)and 1996; Moll, 1997). In addition, permanent endangered least terns (Sterna antillarum), inundation of floodplain wetlands by reser- softshells nest almost exclusively on high voirs has reduced adult and juvenile feeding sandbars, when these habitats are available. habitats for species including Blanding’s tur- Fitch and Plummer, (1975) reported that for tles, Illinois mud turtles, sliders, map turtles, smooth softshells, nearly all activities includ- and western pond turtles (Moll, 1980; Van- ing mating, feeding, and nesting occurred on dewalle and Christiansen, 1996; Reese and or adjacent to sandbars. In Asia, sand min- Welsh, 1998a,b). ing along river banks poses enormous threats Stream and riparian management for freshwater turtles 5 to available nesting habitats of species such of the Missouri River, from 1879 to 1972, as Batagur baska, which faces sand min- water surface area was reduced by 24 654 ha ing in every river it inhabits in at least or 50% of the 1879 surface area primarily five countries (Moll, 1997). In fact, nesting to increase arable land (Funk and Robin- riverine turtles of most species prefer open, son, 1974). Because most freshwater turtle sandy beaches for oviposition (see Table 2) species need the very habitats that have and may seek alternate artificial habitats been reduced (e.g. sandbars, side channels; such as levees if historic nesting habitats are Vandewalle and Christiansen, 1996; Bodie unavailable. et al., 2000; Tucker et al., 1999), it follows that turtle populations may have suffered greater habitat losses (i.e. >50% for the Mis- Human riparian use souri River example). Human uses of riparian areas principally include agriculture, cat- Humans have exerted a particularly strong tle grazing, and urbanization. Urbanization pressure on habitats adjacent to streams. To can directly remove habitat and indirectly use again the example of the 740-km portion change stream hydrology, temperature, and

Table 2. Nest-site timing, habitat type, distance from the stream edge, and data source by freshwater turtle species and location

Species and Nest sites location Timing Habitat Average distance Data source from stream

Flattened musk turtle ( depressus) AL, USA July–September High, sandy 6Ð5 Dodd (1988) shoreline ND1 softshell (Apalone ferox) FL, USA May–August Hammock 22Ð9GoffandGoff clearing ND1 (1935) Map turtle (Graptemys geographica) Quebec, Canada June–August Backwater 2Ð3 (2–3) Gordon and shoreline ND3 MacCulloch (1980) Slider turtle ( scripta) Panama, Panama January–May Open sand 50 (2–320) Moll and Legler and grass ND139 (1971) Smooth softshell (Apalone mutica) KA, USA June–September Open sandbar 38Ð2 (4–90) Fitch and ND104 Plummer (1975) Spiny softshell (Apalone spinifera) AR, USA — Open, sandy 2Ð5 (2–3) Plummer et al. shoreline ND4 (1997) MN, USA June–September Open sandbar 0Ð3 Hedrick and ND1 Holmes (1956) NB, USA June–September Open sandbar 4Ð5 Gehlbach and ND1 Collette (1959) Terecay turtle ( unifilis) Amazonas, Venezuela — High, open 38Ð3 (21–80) Escalona and Fa shoreline ND422 (1998) (Clemmys marmorata) CA, USA June–August clearing 31 Reese (1996) ND1 Mean (š1 SD) 19Ð7š18Ð6m 6 J. R. Bodie

particulate loads (Dunne and Leopold, 1978). Management for monotypic conditions The effects of human riparian use on turtles are manifested in floodplain draining and In many riparian wetlands in the US, flow regulation (discussed above), and pol- management for suites of game species is lution and siltation (discussed below). While common practice (Sparks, 1995). Although occasional flooding of streamside agricultural refuge managers agree that practices to fields may provide habitats functionally anal- encourage nongame species are also a target ogous to historical riparian wetlands (Reese of management (Reid et al., 1989), attempts and Welsh, 1998b; Bodie and Semlitsch, to mimic conditions for all species are unre- 2000a,b), it is clear that conversion of ripar- alistic (Sparks, 1995). Riverine turtles are ian habitats to human land uses is deleteri- adapted to seasonally varying conditions ous to turtle populations (Moll, 1980; Dodd, and may require a multitude of habitats to 1990; Reese and Welsh, 1998b; Bodie and achieve sufficient growth and reproduction. Semlitsch, 2000a,b). For mobile and long-lived species such as turtles, access to habitats with different and annually variable attributes may be espe- cially important for long-term persistence of Pollution and siltation metapopulations (Burke et al., 1995). Plum- mer (1977) found that smooth softshells were Although studied far less than amphib- very mobile, utilizing wide-ranging riverine ians and fishes, freshwater turtles can be habitats. Reese and Welsh (1998b) suggested heavily impacted by pollution and siltation. that a single body of water was not a sufficient Many freshwater turtles have environmental unit of management for western pond turtles, sex determination and are therefore par- and that multiple habitats including streams, ticularly sensitive to endocrine-disrupting shallow marshes and riparian were chemicals that can cause sex reversal or all integral parts of their life histories. Sim- abnormal gonads (Bergeron et al., 1994; Guil- ilarly, Bodie and Semlitsch (2000a,b) found lette and Crain, 1996). Slider turtles are that riverine populations of sliders and false known to incur genetic damage from metal map turtles used at least six general habitats and radioisotope contaminants (Lamb et al., (i.e. agriculture, flooded field, flooded forest, 1995). Ernst et al. (1989) sampled 68 stream river, flood-scoured , slough) within sites in the Black Warrior River drainage in a single year. Likewise, overall turtle species Alabama and determined that agricultural diversity is enhanced with diverse geomor- runoff, surface mining, municipal wastes, and phology within the stream proper (e.g. var- industrial wastes all severely degraded habi- ied substrates, water velocity, water depth; tats of threatened flattened musk turtles. Donner-Wright et al., 1999). in Spain has nearly caused the disappearance of the European pond tur- tle ( orbicularis) from that country’s Unsustainable use rivers (Mascort, 1997). In addition to pol- lutants, Ernst et al. (1989) determined that Freshwater turtles, prized as food, medicinal streams lined with fine silt and clay sup- remedies, and pets, suffer from overuse in ported significantly smaller populations of many parts of the world. Collection of adults flattened musk turtles. From several years and eggs for consumption poses a serious of sampling the Illinois River, Moll (1980) problem in North America, South America, attributed the decline and possible extirpa- India, and Asia (Polisar, 1997; Moll, 1997; tion of smooth softshells and Illinois mud Escalona and Fa, 1998; Gibbons et al., 2000). turtles to continual deposits of silt. Simi- In fact, the commercial trade of riverine lar reasons were given for declines of map species such as softshells has reached crisis turtles and softshells in Missouri (John- proportions in southern China and Vietnam, son, 1992) as well as Kansas (Plummer, with extinction for some species expected 1976). In California, siltation of deep pools within the next decade (Gibbons et al., in the Trinity River has reduced populations 2000). Turtle life-history traits make them of western pond turtles (Reese and Welsh, especially susceptible to changes in large 1998b). juvenile and adult survival (Congdon et al., Stream and riparian management for freshwater turtles 7

1993, 1994). A modest harvest pressure (10% (ND19) and seasonal migrations of individ- per year for 15 years) may result in a 50% uals (ND36) were on average 224 m from reduction in population size (Congdon et al., the stream edge (Table 3). Overwinter sites 1994). When overuse is combined with other for western pond turtles were found up to stresses, it may become even more severe. 423 m from the stream edge, and slider tur- Dodd (1988) and Ernst et al. (1989) report tles migrated up to 1394 m from the edge potential impacts of commercial collectors on during a single year. Under the assumption US populations of threatened flattened musk that the distances turtles migrated from the turtles inhabiting separate river basins in stream were normally distributed (test of nor- Alabama. Close and Seigel (1997) found that mality for log-transformed data in Tables 2 selection by collectors of large individuals and 3; WD0Ð956, PD0Ð670), then the mean might have contributed to significantly larger for all species and movements combined turtles in sites protected from collection (meanD78 m, ND14) represents a distance vs. sites with no protection. This sustained encompassing 50% of the migrations. A ripar- pressure could eventually cause population ian zone encompassing the majority (95% con- declines (see Congdon et al., 1993, 1994). D š Ð aD Ð D ð fidence limits meanp 2 17[ 0 05, df 13] In a similar way, Tucker and Moll (1997) standard deviation/ n) of freshwater turtle suggest that removal of large reproducing migrations would need to encompass a dis- females, even when regulated, may place tance of 150Ð3 m from the stream edge. turtle populations in jeopardy of extinction. The timing of riparian use for most North American freshwater turtles spanned from March through September, with a concentra- Riparian-zone use tion of use for nest deposition and incubation during June through September (Table 2). Data were summarized on 10 species from However, many species have hatchlings that eight US states and four countries (Tables 2 overwintered in the nest during the months and 3). Nests (ND677) of seven freshwater of September through March, and adults of turtle species from North and South Amer- at least one species also overwinter terrestri- ica were found an average of 19Ð7mfrom ally during this time (i.e. western pond tur- the edge of streams (Table 2). Nests were tles; Table 3). Freshwater turtles primarily found as close as 0Ð3 m and up to 320 m chose high, open, sandy habitats for nesting away from the shoreline. Overwinter sites (Table 2). The nesting habitats used included

Table 3. Overwinter-site and seasonal-migration timing, habitat type, average distance from the stream edge, and data source by freshwater turtle species and location

Species and Overwinter sites location Timing Habitat Average distance Data source from stream

Western pond turtle (Clemmys marmorata) CA, USA August–February Forests 168 (39–423) Reese (1996) ND19 Seasonal migrations (Graptemys pseudogeographica) MO, USA March–August Floodplain 353 (0–1133) Bodie and Semlitsch wetlands ND15 (2000a,b) Slider turtle (Trachemys scripta) MO, USA March–August Floodplain 348 (0–1394) Bodie and Semlitsch wetlands ND11 (2000a,b) (Clemmys insculpta) Ontario, Canada May–August Fields and 27 (0–500) Foscarini and forests ND10 Brooks (1993) Mean (š1 SD) 224š157 m 8 J. R. Bodie

sandbars, open shorelines, hammock clear- turtles. Therefore, my suggested 150-m ripar- ings, and levees, occasionally bordered by ian zone should be considered a minimum vegetation. Adult western pond turtles over- estimate pending additional data. wintered in riparian forests. Adult false map How applicable is this recommendation turtles and sliders used a variety of riparian to riparian estimates derived from other wetlands throughout the year (see Bodie and sources? Brosofske et al. (1997) found that Semlitsch, 2000a,b for details). stream microclimate was well maintained by The results indicate freshwater turtles retaining 45-m buffer strips on either side of use large riparian zones to complete sev- streams. A riparian buffer of 60 m may be eral aspects of their life cycle. Elimina- adequate for removal of nonpoint-source pol- tion or alteration of these riparian habi- lutants such as nitrates (Phillips, 1989). In a tats would most likely reduce nest survival similar way, Darveau et al. (1995) determined and, hence, juvenile recruitment into the that 60-m riparian buffers provided habitat breeding population. It would also reduce for many forest-dwelling birds. Hodges and adult survival through lack of overwinter- Krementz, (1996) recommend a 100-m buffer ing and feeding areas, and therefore increase to maintain diversity of the neotropical the risk of extinction for freshwater tur- community. Likewise, Vander-Haegen and tle populations. The importance of riparian Degraaf (1996) suggest that 100-m buffers zones to other myriad fauna including inver- are necessary to reduce edge-related depre- tebrates (e.g. Edwards and Huryn, 1996), dation on bird nests. Davies and Nelson fishes (e.g. Growns et al., 1998), amphibians (1994) suggested that a minimum of 30 m of (e.g. Rudolph and Dickson, 1990), birds (e.g. riparian vegetation was required to maintain Machtans et al., 1996), and mammals (e.g. stream habitat characteristics, invertebrate Thurmond and Miller, 1994) emphasizes the community composition, and fish abundance. universal need for protection of these habi- Rudolph and Dickson (1990) suggested a tats (reviewed by Hall and Lambou, 1990; 30-m buffer for the maintenance of stream- Gregory et al., 1991; Schaefer and Brown, dwelling amphibians, while McComb et al. 1992). (1993) suggested a 100-m buffer provided for Although the extent of riparian area used amphibians and small mammals. Still oth- by freshwater turtles varied due to the pur- ers recommend that protected riparian zones pose of migration, species differences, and should be proportional to stream width, adja- stream and riparian habitat characteristics, cent , and slope (deMaynadier and a generalized riparian zone cannot by defi- Hunter, 1995). The above studies yielded nition encompass every stream and species. mixed recommendations ranging from 30 to However, the suggested 150-m riparian zone 100 m, all less than the biologically deter- encompassing 95% of population migrations, mined 150 m required for 95% of turtle migra- and based on data collected over several tion distances. decades on 10 species, should be robust Perhaps more important than the quantity for management purposes. Furthermore, the of riparian zones is the quality. As mentioned majority of studies concerning turtle nests above, freshwater turtle species such as soft- anecdotally mention estimates of migration shells require habitats very similar to those distances from streams without direct mea- of piping plovers and least terns, birds that surement. In many of these cases, anecdotal have suffered severe declines due to reduc- estimates are much longer than the mea- tions in sandbar nesting and feeding habi- sured distances found in the literature (e.g. tats (US Fish and Wildlife Service, 1985a,b). 400–1600 m, Cagle, 1950; 200 m, Ewert and Stream-management techniques such as sta- Jackson, 1994; 30–1200 m, Tucker et al., bilizing shorelines with riprap and dramati- 1997), suggesting that such data were not cally altering flow not only eliminate habitats collected because of the inherent difficulty spatially but also temporally. Although rela- in measuring long distances across ripar- tively few studies of riparian use by riverine ian habitats. In addition, to my knowledge, turtles have been conducted, it is evident only two studies (i.e. Foscarini and Brooks, that these turtles are mobile and utilize 1997; Bodie and Semlitsch, 2000a,b) have several riparian habitats to complete their purposely measured the size of the riparian life cycle, even within a single year (Jones, area used for seasonal migrations of riverine 1996; Reese and Welsh, 1998b, Bodie and Stream and riparian management for freshwater turtles 9

Semlitsch, 2000a,b). It is clear that more com- Simple passive management (e.g. Galat prehensive data for freshwater turtles are et al., 1998) may accomplish many of these needed, especially on the timing and extent goals. By allowing historical processes to per- of riparian use, and in rapidly developing vade a portion or all of the stream system, parts of the world. the effects of negative management practices (Table 1) may be reduced or eliminated. Ulti- mately, the effectiveness of management for Application to stream and any given stream system is greatly affected riparian management by surrounding systems and should be adapt- able to enlightened approaches (Christensen et al., 1996). True to the ecosystem manage- Contemporary managers are faced with ment approach, the goals of turtle population the daunting task of ecosystem manage- monitoring should ensure successful repro- ment—management driven by adaptability duction, juvenile recruitment, and a diverse through monitoring, and based on sound riverine species assemblage. In addition, knowledge of ecological processes that sus- many single riverine species display strong tain ecosystem diversity and function (Chris- genetic distinctiveness among drainages, sig- tensen et al., 1996). It is therefore essential nifying the drainage as the proper manage- that biologists emphasize the processes that ment unit for preserving genetic resources sustain important, yet often neglected, faunal important for long-term species survival components of stream system diversity such (Roman et al., 1999). as freshwater turtles. Based on the literature, I encourage these practices within a min- I offer the following recommendations. imum 150-m riparian zone, an area that A key hydrological process is the connec- should be considered biologically critical to tion of stream and riparian habitats through freshwater turtles. A buffer as originally groundwater and over- flow. This con- intended (i.e. Schonewald-Cox, 1988) would nection provides riparian habitat complexity extend farther than the 150-m core habitat to and woody debris through hydraulic reduce potentially damaging edge effects (see and deposition. For regulated streams, the Semlitsch, 1998). I also encourage additional timing of connection should be adapted to documentation of turtle riparian habitat use, the historical hydroperiod, providing con- especially for species of federal or interna- ditions that favor formation and mainte- tional conservation concern or with special nance of temporary wetlands and high, sandy habitat requirements (e.g. terrestrial over- beaches throughout the warm months. Like- wintering). wise, it is crucial to reduce or eliminate artificial changes in streamflow throughout the cool months when hatchlings, juveniles, Acknowledgements and adults overwinter. Mechanical and other disturbance of riparian nesting and overwin- I thank D. Galat, W. Gibbons and D. Moll for tering habitats should be disallowed through- insightful manuscript criticisms. Much of the out the year. Monitoring, a tenet of ecosystem tenor of this paper was inspired by discus- sions with R. Semlitsch, who also critiqued the management, is especially critical for riverine manuscript. This review was partially supported turtle populations that experience collection by cooperative research grant MCC 82-01-00-136 or consumption, as effects of overuse of these to R. Semlitsch between the University of Missouri long-lived species may take years to detect and the Missouri Department of Conservation, and and overcome. Indeed, the overexploitation of in part with Federal Aid in Wildlife Restoration Act funds under Missouri’s Pittman-Robertson Asian riverine turtles can only benefit from a Project W-13-R. moratorium on collection. To sustain species diversity, broad-scale management for condi- tions that promote permanent lentic habitats References should be avoided in lotic systems. Finally, sources of and silt entering stream Benke, A. C., Henry III, R. L., Gillespie, D. M. and systems must be identified and eliminated or Hunter, R. J. (1985). Importance of snag habitat managed through, for example, incorporation for production in southeastern streams. of vegetative buffers. Fisheries 10, 8–13. 10 J. R. Bodie

Bergeron, J. M., Crews, D. and McLachlan, J. A. Journal of Marine and Freshwater Research 45, (1994). PCBs as environmental estrogens: turtle 1289–1305. sex determination as a biomarker of environ- deMaynadier, P. G. and Hunter, M. L., Jr (1995). mental contamination. Environmental Health The relationship between forest management Perspectives 102, 780–781. and amphibian ecology: a review of the North Bodie, J. R. and Semlitsch, R. D. (2000a). Spatial American literature. Environmental Reviews 3, and temporal use of floodplain habitats by lentic 230–261. and lotic species of aquatic turtles. Oecologia van Dijk, P. P. and Thirakhupt, K. (1996). South- 122, 138–146. east Asian Chitra: from distinction to extinction Bodie, J. R. and Semlitsch, R. D. (2000b). Size- in 15 years? In Proceedings of the International specific mortality and natural selection in fresh- Congress of Chelonian Conservation. pp. 62–63. water turtles. Copeia 2000, 732–739. Chelonian Research Foundation, Lunenburg, Bodie, J. R., Semlitsch, R. D. and Renken, R. B. Massachusetts. (2000). Diversity and structure of turtle assem- Dodd, C. K., Jr (1988). Disease and population blages: associations with wetland characters declines in the flattened musk turtle Ster- across a floodplain landscape. Ecography 23, notherus depressus. The American Midland Nat- 444–456. uralist 119, 394–401. Brosofske, K. D., Chen, J., Naiman, R. J. and Dodd, C. K., Jr (1990). Effects of habitat frag- Fraklin, J. F. (1997). Harvesting effects on mentation on a stream-dwelling species, the microclimatic gradients from small streams flattened musk turtle Sternotherus depressus. to uplands in western Washington. Ecological Biological Conservation 54, 33–45. Applications 7, 1188–1200. Donner-Wright, D. M., Bozek, M. A., Probst, J. R. Burke, V. J. and Gibbons, J. W. (1995). Terrestrial and Anderson, E. M. (1999). Responses of tur- buffer zones and wetland conservation: a case tle assemblage to environmental gradients in study of freshwater turtles in Carolina bay. the St. Croix River in Minnesota and Wiscon- 9, 1365–1369. sin, U.S.A. Canadian Journal of Zoology 77, Burke, V. J., Greene, J. L. and Gibbons, J. W. 989–7000. (1995). The effect of sample size and study Dunne, T. and Leopold, L. B. (1978). Water in duration on metapopulation estimates for slider Environmental Planning. San Francisco, CO: turtles (Trachemys scripta). Herpetologica 51, W. H. Freeman. 451–456. Edwards, E. D. and Huryn, A. D. (1996). Effect of Cagle, F. R. (1950). The life history of the slider riparian land use on contributions of terrestrial turtle, scripta troosti (Holbrook). invertebrates to streams. Hydrobiologia 337, Ecological Monographs 20, 31–54. 151–159. Christensen, N. L., Bartuska, A. M., Brown, J. H., Ernst,C.H.,Cox,W.A.andMarion,K.R. Carpenter, S., D’Antonio, C., Francis, R. et al. (1989). The distribution and status of the (1996). The report of the Ecological Society flattened musk turtle, Sternotherus depressus of America committee on the scientific basis (Testudines: ). Tulane Studies in for ecosystem management. Ecological Applica- Zoology and Botany 27, 1–20. tions 6, 665–691. Escalona, T. and Fa, J. E. (1998). Survival of nests Close, L. M. and Seigel, R. A. (1997). Differences of the terecay turtle (Podocnemis unifilis)inthe in body size among populations of red-eared Nichare-Tawadu Rivers, Venezuela. Journal of sliders (Trachemys scripta elegans) subjected Zoology 244, 303–312. to different levels of harvesting. Chelonian Ewert, M. A. and Jackson, D. R. (1994). Nest- Conservation and Biology 2, 563–566. ing Ecology of the Alligator Snapping Tur- Congdon, J. C., Dunham, A. E. and van Loben tle (Macroclemys temminckii) along the Lower Sels, R. C. (1993). Delayed sexual maturity and Apalachicola River, Florida. Florida Game demographics of Blanding’s turtles (Emydoidea and Freshwater Fish Commission, Non-game blandingii): implications for conservation and Wildlife Program Final Report, Tallahassee, management of long-lived organisms. Conserva- Florida, USA. tion Biology 7, 826–833. Fitch, H. S. and Plummer, M. V. (1975). A Congdon, J. C., Dunham, A. E. and van Loben preliminary ecological study of the soft-shelled Sels, R. C. (1994). Demographics of common turtle muticus in the Kansas River. snapping turtles ( serpentina): impli- Israel Journal of Zoology 24, 28–42. cations for conservation and management of Foscarini, D. A. and Brooks, R. J. (1993). A pro- long-lived organisms. American Zoologist 34, posal to standardize data collection and impli- 397–408. cations for management of the wood turtle, Darveau, M., Beauchesne, P., Belanger, L., Clemmys insculpta, and other freshwater tur- Huot, J. and Larue, L. (1995). Riparian forest tles in Ontario, Canada. In Proceedings of the strips as habitat for breeding birds in boreal for- Conservation, Restoration, and Management of est. Journal of Wildlife Management 59, 67–78. and Turtles: An International Confer- Davies, P. E. and Nelson, M. (1994). Relationships ence. pp. 203–209. State University of New York, between riparian buffer widths and the effects of Purchase, New York. on stream habitat, invertebrate commu- Fuselier, L. and Edds, D. (1994). Habitat parti- nity composition and fish abundance. Australian tioning among three sympatric species of map Stream and riparian management for freshwater turtles 11

turtles, Graptemys. Journal of Herpetol- Johnson, T. R. 1992. The Amphibians and Rep- ogy 28, 154–158. tiles of Missouri. Jefferson City, MO: Missouri Galat, D. L., Fredrickson, L. H., Humburg, D. D., Department of Conservation. Bataille, K. J., Bodie, J. R., Dohrenwend, J. Jones, R. L. (1996). Home ranges and seasonal et al. (1998). Flooding to restore connectivity of movements of the turtle Graptemys flavimacu- regulated, large-river wetlands. BioScience 48, lata. Journal of Herpetology 30, 376–385. 721–733. Kushlan, J. A. and Jacobsen, T. (1990). Environ- Gehlbach, F. R. and Collette, B. B. (1959). Distri- mental variability and reproductive success of butional and biological notes on the Nebraska Everglades alligators. Journal of Herpetology herpetofauna. Herpetologica 15, 141–173. 24, 176–184. Gibbons, J. W., Scott, D. E., Ryan, T., Buhlmann, Lamb et al. (1995). K., Tuberville, T., Greene, J. et al. (2000). The Lindeman (1999). global decline of reptiles, dej´ a` vu amphibians. Machtans, C. S., Villard, M. A. and Hannon, S. J. BioScience 50, 653–666. (1996). Use of riparian buffer strips as move- Goff, D. S. and Goff, C. C. (1935). On the incubation ment corridors by forest birds. Conservation of a clutch of eggs of ferox (Schneider). Biology 10, 1366–1379. Copeia 1935, 156. Marzolf, G. R. (1978). The Potential Effects of Gordon, D. M. and MacCulloch, R. D. (1980). Clearing and Snagging on Stream Ecosystems. An investigation of the ecology of the map US Fish and Wildlife Service, Office of Biological turtle, Graptemys geographica (LeSeur), in the Services, Washington, DC. FWS/OBS- 78/14. northern part of its range. Canadian Journal of Mascort, R. (1997). An overview of a threat- Zoology 58, 2210–2219. ened population of the , Gore, J. A. and Shields, F. D., Jr (1995). Can large Emys orbicularis.InProceedings of the Con- rivers be restored? BioScience 45, 142–152. servation, Restoration, and Management of Tor- Graham, T. E. and Graham, A. A. (1997). Ecology toises and Turtles: An International Conference of the eastern spiny softshell, Apalone spinifera (J. V. Abbema, ed.), pp. 441–443. New York spinifera, in the Lamoille River, Vermont. Che- Turtle and Society, New York. lonian Conservation and Biology 2, 363–369. McComb, W. C., McGarigal, K. and Anthony, R. G. Gregory, S. V., Swanson, F. J., McKee, W. A. (1993). Small mammal and amphibian abun- and Cummins, K. W. (1991). An ecosystem dance in streamside and upslope habitats of perspective of riparian zones. BioScience 41, mature Douglas-fir stands, western Oregon. 540–551. Northwest Scientist 67, 7–15. Growns, I. O., Pollard, D. A. and Gherke, P. C. Moll, D. (1980). Dirty River Turtles. Natural (1998). Changes in river fish assemblages asso- History 89, 42–49. ciated with vegetated and degraded banks, Moll, D. (1990). Population sizes and foraging upstream of and within nutrient-enriched zones. ecology in a tropical freshwater stream turtle Fisheries Management and Ecology 5, 55–69. community. Journal of Herpetology 24, 48–53. Guillette, L. J. and Crain, D. A. (1996). Endocrine- Moll, E. O. (1997). Effects of habitat alteration on disrupting contaminants and reproductive abn- river turtles of tropical Asia with emphasis on ormalities in reptiles. Comments in Toxicology sand mining and dams. In Proceedings of the 56, 381–399. Conservation, Restoration, and Management of Haeuber, R. A. and Michener, W. K. (1998). Policy Tortoises and Turtles: An International Confer- implications of recent natural and managed ence (J. V. Abbema, ed.), pp. 37–41. New York floods. BioScience 48, 765–772. Turtle and Tortoise Society, New York. Hall, H. D. and Lambou, V. W. (1990). The ecolog- Moll, E. O. and Legler, J. M. (1971). The life ical significance to fisheries of bottomland hard- history of a neotropical slider turtle, Pseudemys wood systems: values, detrimental impacts, and scripta (Schoepff), in Panama. Bulletin of the assessment: the report of the fisheries group. Los Angeles County Museum of Natural History Ecological Processes and Cumulative Impacts Science 11, 1–102. (J. G. Gosselink, L. C. Lee and T. A. Muir, eds), Molles, M. C., Jr, Crawford, C. S., Ellis, L. M., pp. 481–531. Chelsea, MI: Lewis Publishers. Valett, H. M. and Dahm, C. N. (1998). Managed Hedrick, R. M. and Holmes, J. C. (1956). Addi- flooding for riparian ecosystem restoration. Bio- tional Minnesota herpetological notes. Flicker Science 48, 749–756. 28, 123–126. Moyle, P. B. and Vondracek, B. (1985). Persistence Hesse, L. W. and Mestl, G. E. (1993). An alterna- and structure of the fish assemblage in a small tive hydrograph for the Missouri River based on California stream. Ecology 66, 1–13. the precontrol condition. North American Jour- Naiman et al. (2000). nal of Fisheries Management 13, 360–366. Petts, G. E. (1984). Impounded Rivers. Chichester, Hirsch, R. M., Miller, T. L. and Hamilton, P. A. UK:JohnWiley&Sons. (2001). Using today’s science to plan for tomor- Phillips, J. D. (1989). Nonpoint source pollution row’s water policies. Environment 43, 8–17. control effectiveness of riparian forests along a Hodges, M. F., Jr and Krementz, D. G. (1996). coastal plain river. Journal of Hydrology 110, Neotropical migratory breeding bird communi- 221–238. ties in riparian forests of different widths along Plummer, M. V. (1976). Some aspects of nesting the Altamaha River, Georgia. Wilson Bulletin success in the turtle, Trionyx muticus. Herpeto- 108, 496–506. logica 32, 353–359. 12 J. R. Bodie

Plummer, M. V. (1977). Activity, habitat and Semlitsch, R. D. (1998). Biological delineation of population structure in the turtle, Trionyx terrestrial buffer zones for pond-breeding sala- muticus. Copeia 1977, 431–440. manders. Conservation Biology 12, 1113–1119. Plummer, M. V., Mills, N. E. and Allen, S. L. Simpson, P. W., Newman, J. R., Keirn, M. A., (1997). Activity, habitat, and movement pat- Matter, R. M. and Guthrie, P. A. (1982). Manual terns of softshell turtles (Trionyx muticus)in of Stream Channelization Impacts on Fish and a small stream. Chelonian Conservation and Wildlife. US Fish and Wildlife Service, Office of Biology 2, 514–520. Biological Services, Washington, DC. FWS/OBS- Polisar, J. (1997). Effects of exploitation on Der- 82/24. matemys mawii populations in northern Belize Sparks, R. E. (1995). Need for ecosystem man- and conservation strategies for rural river- agement of large rivers and their floodplains. side villages. In Proceedings of the Conser- BioScience 45, 168–182. vation, Restoration, and Management of Tor- Sparks, R. E., Nelson, J. C. and Yin, Y. (1998). toises and Turtles: An International Conference Naturalization of the flood regime in regulated rivers. BioScience 48, 706–720. (J. V. Abbema, ed.), pp. 441–443. New York Statistical Abstract of the US (1998). US Wash- Turtle and Tortoise Society, New York. ington, DC: Census Bureau. Reese, D. A. (1996). Comparative Demography Thurmond, D. P. and Miller, K. V. (1994). Small and Habitat Use of Western Pond Turtles in mammal communities in streamside manage- Northern California: The Effects of Damming ment zones. Brimleyana 1994, 125–130. and Related Alterations. PhD Thesis. University Timken, R. L. (1968). Graptemys pseudogeograph- of California, Berkeley, CA, USA. ica in the upper Missouri River of the north- Reese, D. A. and Welsh, H. H., Jr (1998a). Com- central United States. Journal of Herpetology 1, parative demography of Clemmys marmorata 1–4. populations in the Trinity River of California in Tinkle, D. W. (1958). The systematics and ecology the context of dam-induced alterations. Journal of the Sternothaerus carinatus complex (Testu- of Herpetology 32, 505–515. dinata: ). Tulane Studies in Zoology Reese, D. A. and Welsh, H. H., Jr (1998b). Habitat 6, 3–56. use by western pond turtles in the Trinity River, Toth, L. A., Melvin, S. L., Arrington, D. A. and California. Journal of Wildlife Management 62, Chamberlain, J. (1998). Hydrologic manipula- 842–853. tions of the channelized Kissimmee River. Bio- Reid, F. A., Kelley, J. R., Jr, Taylor, T. S. and Science 48, 757–764. Fredrickson, L. H. (1989). Upper Mississippi val- Tucker, J. K. and Moll, D. (1997). Growth, repro- ley wetlands: refuges and moist-soil impound- duction, survivorship in the red-eared turtle, ments. In Habitat Management for Migrating Trachemys scripta elegans, in Illinois, with con- and Wintering Waterfowl in North America servation implications. Chelonian Conservation (R. L. Pederson and R. M. Kaminski, eds), and Biology 2, 352–357. pp. 181–202. Texas Tech University Press, Lub- Tucker, J. K., Janzen, F. J. and Paukstis, G. L. bock, Texas, USA. (1997). Responses of embryos of the red-eared Roman, J., Santhuff, S. D., Moler, P. E. and turtle (Trachemys scripta elegans)toexperi- Bowen, B. W. (1999). Population structure and mental exposure to water-saturated substrates. cryptic evolutionary units in the alligator snap- Chelonian Conservation and Biology 2, 345–351. ping turtle. Conservation Biology 13, 135–142. Tucker, J. K., Gritters, S. A. and Hrabic, R. A. Rudolph, D. C. and Dickson, J. G. (1990). (1999). Turtle Communities: An Analysis of Streamside zone width and amphibian and rep- Aquatic Species in the Upper tile abundance. Southwestern Naturalist 35, System. Onalaska, Washington: Environmental 472–476. Management Technical Center. US Army Corps of Engineers (2000). Civil works Schaefer, J. M. and Brown, M. T. (1992). Designing program statistics. Information Paper CECW- and protecting river corridors for wildlife. Rivers ZP, 6 April 2000. 3, 14–26. US Fish and Wildlife Service (1985a). Deter- Schlosser, I. J. (1990). Environmental variation, mination of endangered and threatened sta- life history attributes, and community structure tus for the piping plover. Federal Register 50, n stream fishes: implications for environmental 50720–50734. management and assessment. Environmental US Fish and Wildlife Service (1985b). Interior Management 14, 621–628. population of the least tern determined to be Schonewald-Cox, C. M. (1988). Boundaries in the endangered. Federal Register 50, 21784–21792. protection of nature reserves. BioScience 38, US Fish and Wildlife Service (1987). Determina- 480–486. tion of threatened status for the flattened musk Sedell, J. R. and Froggatt, J. L. (1984). Impor- turtle (Sternotherus depressus). Federal Register tance of streamside forests to large rivers: the 50, 45638–45643. isolation of the Willamette River, Oregon, USA, US Fish and Wildlife Service (2000). Endangered from its floodplain by snagging and streamside and threatened wildlife and plants. Code of forest removal. Verhandlungen Internationale Federal Regulations 50CFR17. 11, 102–143. Vereinigung fur Theoretische und Augewandte Funk, J. L. and Robinson, J. W. (1974). Changes Limnologie 22, 1828–1834. in the Channel of the Lower Missouri River Stream and riparian management for freshwater turtles 13

and Effects on Fish and Wildlife. Missouri Vannote, R. L., Minshall, G. W., Cummins, K. W., Department of Conservation, Jefferson City, Sedell, J. R. and Cushing, C. E. (1980). The Missouri, USA. river continuum concept. Canadian Journal of Vandewalle, T. J. and Christiansen, J. L. (1996). Fisheries and Aquatic Sciences 37, 130–137. A relationship between river modification and Vogt, R. C. and Villareal, J. L. B. (1993). Species species richness of freshwater turtles in Iowa. abundance and biomass distribution in fresh- Journal of the Iowa Academy of Sciences 103, water turtles. In Proceedings of the Conser- 1–8. vation, Restoration and Management of Tor- Vander-Haegen, W. M. and Degraaf, R. M. (1996). toises and Turtles: An International Conference Predation on artificial nests in forested riparian pp. 210–218. State University of New York, Pur- buffer strips. Journal of Wildlife Management chase, New York. 60, 542–550. 14 J. R. Bodie

Author Queries

Manuscript Page Query Number

References Statistical Abstract of the US Author? Title? Please clarify. Tables Table 3 Bodie and Semlitsch (2000) (a) or (b) ð2? msp 3 (9 up) Naiman et al (2000) Please provide a full reference. msp 6 (8 down) Lindeman (1999) Please provide a full reference msp 7 (1 down) Bodie and Semlitsch (2000) (a) or (b)? msp 7 (7 down) Bodie and Semlitsch (2000) (a) or (b)? msp 8 (1 down) Bodie and Semlitsch (2000) (a) or (b)? msp 9 (9 down) Bodie and Semlitsch (2000) (a) or (b)? msp 10 (10 up) Bodie and Semlitsch (2000) (a) or (b)? msp 10 (8 up) Bodie and Semlitsch (2000) (a) or (b)? msp 10 (2 up) Lamb et al (1995) Please provide a full reference msp 11 (1 up) Bodie and Semlitsch (2000) (a) or (b)? msp 13 (1 up) Bodie and Semlitsch (2000) (a) or (b)? msp 14 (3 up) Bodie and Semlitsch (2000) (a) or (b)? msp 16 (2 down) Bodie and Semlitsch (2000) (a) or (b)?