2008 NORTHEASTERN NATURALIST 15(3):349–362 Stepwise Increases in Maximum Prey Size of Larval Creek Chubs, Semotilus atromaculatus, in an Urbanized Ohio Stream Stacey A. Ward1 and Miles M. Coburn1,* Abstract - The diet of larval Semotilus atromaculatus (Creek Chubs) was examined in fish collected from an urbanized stream with a limited food base. Chironomids comprised nearly 90% of food items. They appeared in the gut of early larvae and continued to be the main food source as size increased; cladocerans were the second most abundant food item. Both average and maximum prey size were examined. Overall, average prey size increases significantly with standard length (SL). Maximum prey size is gape-limited twice, from the early to mid-mesolarval stage and again in the late mesolarval stage. Significant differences were observed in the maximum size of chironomids ingested among fish of four size ranges, less than8.0 mm SL, 8.0–9.6 mm SL, 9.7–14.1 mm SL, and >14.1 mm SL, indicating maximum prey size increases rapidly at 8.0 mm SL, 9.6 mm SL, and again at 14.1 mm SL, yet maximum prey size within each group remains constant. The degree of cranial ossification and fin development at these break points was examined with cleared and double-stained specimens. For fishes less than 8.0 mm SL, ossification is just commencing, and maximum prey size is gape limited. At 9.6 mm SL, ossification appeared nearly complete in the caudal fin and in key bones involved in food capture and processing. There was no obvious correlation between ossification of the skeleton and the third break point at 14.1 mm SL. The results from this study suggest that ossification of feeding apparatus and caudal fin may play an important role in the ability of Creek Chubs to capture larger prey in the mesolarval stage at about 9.6 mm SL, but other factors likely account for the increase in prey size at about 14 mm SL. Introduction The larval stage of fish is a vulnerable and critical time (Gerking 1994, Lammens and Hoogenboezem 1991), and much effort has been directed towards understanding factors that affect larval feeding success (Bremigan and Stein 1994, Sanderson and Kupferberg 1999). Until they reach a certain length, often about 8–10 mm, larval fish are gape-limited predators (DeVries et al. 1998), with mouth size being the decisive factor in prey selection (Cunha and Planas 1999). Krebs and Turingan (2003) have noted that as larvae grow the average size of prey in many species is less than that predicted by gape:prey size relationships. Others have identified a host of intrinsic and extrinsic factors related to such things as visual acuity, development of larval head morphology, and prey shape, size, abundance, and mobility that play a role in successful larval feeding (Bremigan and Stein 1994, Bremigan et al. 2003, Cunha and Planas 1999). This study examined the relationship 1Department of Biology, John Carroll University, 20700 North Park Boulevard, University Heights, OH 44118. *Corresponding author - coburn@jcu.edu. 350 Northeastern Naturalist Vol. 15, No. 3 among prey size, body size, gape, and ossification of the skeletal system in larval Semotilus atromaculatus (Mitchill) (Creek Chub), feeding in an urbanized stream with an abundant food base of very limited diversity. Adult Creek Chubs are best described as opportunistic generalist feeders, preying on whatever organisms that are available, from the surface drift to the benthos and plant material (Copes 1978, Dinsmore 1962, Quist et al. 2006). Dinsmore (1962) found that Creek Chubs were adaptable in their food habits with the changing conditions of an Iowa river. Rosati et al. (2003) and Sellman et al. (2002) found Creek Chubs to be representative samplers of diatoms in northeastern Ohio streams. Even though they are variable in their food habits, Creek Chubs are generally insectivorous, feeding at all depths of the stream on chironomid larvae, mayfly nymphs, and molluscs, along with terrestrial insects that have accidentally fallen or landed in the stream and even fast-moving fish (Barber and Minckley 1971). Magnan and FitzGerald (1984) found juvenile Creek Chubs in shallow water feeding on relatively small prey, such as adult dipterans and aquatic adult coleopterans. Other studies have found that young juveniles feed upon aquatic and terrestrial insects and amphipods (Barber and Minckley 1971, Copes 1978, McMahon 1982), but little is known about what food sources Creek Chubs are utilizing between hatching and the juvenile stage. Doan Brook, an urbanized stream in Cuyahoga County, OH, has only three resident fish species—Lepomis cyanellus Rafinesque (Green Sunfish), Rhinichthys atratulus (Hermann) (Blacknose Dace), and Creek Chub. The food base consists largely of chironomids, cladocerans, algae, diatoms, and terrestrial invertebrates. Benthic sampling by the Northeast Ohio Regional Sewer District (1992) at the location of our study site in the late 1980s and 1990s yielded only eight benthic macroinvertebrate taxa. The limitation imposed by Doan Brook’s restricted food base affords an opportunity to focus not only on diet switching, but also to look at size preferences of the same prey item ingested by larval Creek Chubs. The gut contents of Creek Chubs from early swimup mesolarvae through the early juvenile stage were examined to determine: 1) the sequence of food-types consumed by young fish as they grow, identifying at what size and in which developmental stage they first begin feeding on chironomids, the main food source of juvenile and adult Chubs in Doan Brook; 2) at what size gape no longer limits prey size; 3) the relationship between body size and maximum prey size; and to investigate 4) the correlation between maximum prey size and the ossification of the feeding apparatus and fin rays. Materials and Methods During the 2005 sampling period (28 May to 25 Aug 2005), 1026 specimens of Creek Chub mesolarvae through early juvenile stages were collected in Doan Brook, a small urbanized stream located within Cuyahoga County, OH, at a site about 0.5 km upstream from The Nature Center at Shaker 2008 S.A. Ward and M.M. Coburn 351 Lakes (41°29'04"N, 81°34'14"W). Identification of Creek Chub gravel nests was performed before sampling took place. Beginning in April, nests were searched for the presence of eggs. Once eggs were found, the shallow areas and riffles were searched for the presence of early larvae. Collection of larvae occurred from 10 am to 2 pm, using aquarium nets, dip nets, and a seine. Samples were taken every 3 days, switching to once per week during July, with a final sample taken in late August. Collected fish were transported in buckets to John Carroll University, euthanized within an hour of capture, fixed overnight in 10% neutral-buffered formalin, and transferred through an alcohol series to 70% EtOH. The standard length (SL) of each specimen was determined using an ocular microscope mounted on a Nikon SMZ-10 dissecting scope. Larval fishes were classified as meso- or metalarvae based on criteria identified by Snyder et al. (1977). No protolarvae were captured in this study. Mesolarvae possessed at least one but not the full complement of distinct principal rays in the median fins, and the pelvic buds were not yet apparent. Metalarvae possessed a full complement of distinct principal rays in the median fins, and possessed pelvic buds or fins. Juveniles possessed a full complement of distinct, segmented rays in all fins. Based on examination of preserved, cleared, and stained Creek Chubs, the meso- to metalarval transition occurs between 10.65 mm SL (pelvic bud present in a few specimens) and 10.95 mm SL (pelvic bud present in all specimens). Segmentation of the innermost pelvic and pectoral fin rays was visible in specimens between 19.0 mm and 20.5 mm SL, marking the transition from metalarvae to juveniles. The intestines of 152 fish ranging in size from 6.5–31.65 mm SL were dissected and the contents placed on microscope coverslips containing a drop of Taft’s syrup medium (TSM); this material was dried down and mounted on microscope slides. The coverslips were sealed to the slides with clear nail polish (Environmental Protection Agency TSM protocol, http://www.epa.gov/owow/monitoring/rbp/ch06main.html). The larval fish in this study did not masticate their prey and the exoskeletons of chironomid heads and cladoceran bodies passed through the gut nearly intact. The gut contents were analyzed, and the size of prey items measured using Olympus MicroSuiteTM Basic Edition software and a digital camera mounted on an Olympus IX71 inverted microscope. For chironomids, maximum head width was measured; for cladocerans, the maximum width of the thorax in lateral view was determined. Prey width has been shown to be a more important consideration than length in determining maximum prey size since fish larvae mostly swallow their prey head first (Cunha and Planas1999, Hjelm et al. 2003). Gape was measured on a separate set of 20 preserved fish ranging from 6.9–39.95 mm SL. Angle and gape measurements were made from digital photographs obtained from a camera mounted on an Olympus SZ11 dissecting scope. A rod with a conical tip was inserted into the mouth until the gape 352 Northeastern Naturalist Vol. 15, No. 3 opened to a 900 angle (Qin and Hillier 2000). Gape was measured from the anterior tip of the premaxilla to the anterior tip of the dentary. An additional 25 specimens, ranging in size from 6.2–16.2 mm SL, were cleared and double-stained (Potthoff 1984) to determine accurately the transition points between larval stages and to examine relationship between prey size and the degree of cranial ossification and fin development. Mabee et al. (1998) have shown that some shrinkage, about 3% or 0.3 mm for a 10 mm specimen, of larval Tilapia mossambica (Peters) (Tilapia) of similar sizes to our specimens occurs during the preservation, clearing, and staining process, although variation among individuals can be great (95% CI = ±4.60 mm for a 10 mm fish). Our cleared and stained specimens were measured only once after they were prepared and we presume some shrinkage occurred relative to preserved specimens. Specimens were stored in a 50% glycerin/Alizarin Red solution to prevent stain leaching from bones. During the course of development, the ability to ingest a larger prey item could be due to morphological changes such as gape size, ossification of cranial elements, fin development, or greater functionality of the sensory system (Kawakami and Tachihara 2005, Krebs and Turingan 2003, Makrakis et al. 2005, Reyes-Marchant et al. 1992). With the specimens available, it was feasible to look only for correlations between prey size and the degree of ossification in bones involved with feeding or fin development as the fish grow. Each bone involved in feeding was scored according to its degree of ossification: none (if the bone was unossified), early ossification (stain visible on a small portion of the bone), mid-ossification (stain not extending through the entire bone), or late ossification (entire bone stained). Bones that were scored included the premaxilla, maxilla, dentary, anguloarticular, ceratohyals, ceratobranchials of gill arches 1–4, ceratobranchial 5 (pharyngeal arch), palatine, pterygoids, hyomandibula, quadrate, opercle, subopercle, branchiostegal rays, and the cleithrum. Fin development was examined in the caudal, pectoral, dorsal, anal, and pelvic fins by noting if the fin rays had formed and the degree of ossification present. SPSS was used for statistical analyses [SPSS 2003]. Fish were placed into four putative groups based on initial inspection of feeding data. The data were log-transformed since distribution of prey sizes within each group was non-normal. We recognize that, at the time of capture, many fish would not have consumed chironomids or cladocerans as large as they were capable of eating. Since we were interested in determining the maximum size of prey a larval fish can consume, we treated the data as follows. All chironomids consumed by fish in each group were pooled across dates, and the upper 10% of chironomid head widths was determined for that group. We retained for analysis only fish that had consumed chironomids falling into the upper 10% (9 of 15 fish were retained in group 1; 10 of 25 fish in group 2; 26 of 51 fish in group 3; and 18 of 39 fish in group 4). After determining the largest single chironomid consumed by each of the remaining fish, a single-factor ANOVA compared those chironomids among fish size groups. 2008 S.A. Ward and M.M. Coburn 353 Following a significant F-ratio (α = 0.05), Tukey-Kramer tests were used to make pairwise comparisons between means. Using the same methodology, we analyzed the gut contents of 24 fish from one collection date to address the question of whether fish were ingesting larger chironomids from a wide range of available sizes or were consuming large chironomids because they were the only prey available. We also compared maximum size of cladocerans consumed using a similar approach except that we included, because of the narrower size range of cladocerans and smaller sample sizes, all fish that had consumed cladocerans in the upper 50% percentile of size rather than just the upper 10%. Results The gut contents of 152 dissected fish revealed 129 individuals with chironomids in their intestines; of those, 60 individuals also had consumed cladocerans. Five fish had eaten only cladocerans; 16 fish had other prey items (plant material, unidentifiable insects or ants); and the intestines of two fish were empty. On occasion, other food items included small numbers of copepods, diatoms, snails, plant material, and terrestrial insects. A total of 1110 prey items was measured, with 996 (89.7%) being chironomids; the average fish intestine contained 7.9 chironomids and 2.3 cladocerans. No appreciable switching of prey items occurred in Doan Brook Creek Chubs as they grew. Chironomids appeared in the gut of the smallest larva and continued to be the main food source in larger fish. A regression of mean chironomid head width against SL of the fish which consumed them showed that mean prey size changes significantly with increasing body length (R2 = 0.117, df = 83, slope = 0.0032, F = 11.01, P < 0.001), with the low R2 value indicating that Creek Chub larvae opportunistically consume a variety of prey sizes. A close inspection of the data suggested that a non-linear stepwise relationship between predator size and maximum prey size may exist at three points in the meso- and metalarval stages. Breaks in maximum prey size of Creek Chubs appear to occur at about 8.0 mm, 9.6 mm, and 14.1 mm SL (Fig. 1), thereby creating four groups of fish: 1) those <8.0 mm SL, 2) those from 8.0–9.5 mm SL, 3) those from 9.6–14.1 mm SL, and 4) those >14.1 mm. Gape, as calculated from preserved fish, increases linearly with standard length (Fig. 1, dotted line). When compared to the maximum size of chironomids consumed, gape appears to be a limiting factor twice, once when mesolarvae are smaller than 8.0 mm SL and again for mesolarvae of about 9.6 mm SL. For mesolarvae up to 8.0 mm SL, widest chironomid head widths are equal to or even exceed measured gape and gape is clearly limiting (Fig. 1). For larvae between 8.0 and 9.6 mm SL the largest chironomids consumed fall below the size predicted by gape measurements, but at 9.6 mm SL, maximum prey size increases rapidly and gape may again be limiting (Fig. 1). Maximum prey size increases again when metalarvae reach approximately 14 mm SL, but at this size gape is no longer limiting. 354 Northeastern Naturalist Vol. 15, No. 3 A single-factor ANOVA comparing fish in the four size groups which had consumed large chironomids yielded a significant F-ratio (F(3, 59) = 69.63, P < 0.001) indicating differences among group means. Tukey-Kramer tests showed a significant difference in each pairwise comparison between means (Table 1). From this analysis, we conclude that maximum prey size increases rapidly between groups but remains fairly constant within each group. It is possible that growing Creek Chub larvae were consuming larger chironomids because those may have been the only prey available to them. To test this, we examined 17 fish (8.40–16.50 mm SL), collected on the same date, which had ingested large chironomids in the upper 10% of all chironomids consumed during the study. We tested four group-2 fish, nine group-3 fish, and four group-4 fish; no group-1 fish were sampled for this date. The head widths of chironomids ingested by fish in this sample varied by a factor of 10, from 0.057 mm–0.599 mm, a range comparable to 12.5-fold range (0.049–0.617 mm) of all chironomids measured during the study. As in the larger analysis, an ANOVA of the upper 10% of chironomids consumed found a significant F-statistic (F(2, 14) = 41.78, P < 0.001), and Tukey-Kramer Figure 1. Equation and regression line of gape (dotted line) compared to mean (solid circles) and maximum-minimum range (vertical lines) of head width measurements of all chironomids found in Creek Chub larvae up to 20 mm SL. Arrows indicate breaks at 8.0 mm SL, 9.6 mm SL, and 14.1 mm SL where the maximum size of ingested chironomids increases significantly. 2008 S.A. Ward and M.M. Coburn 355 tests revealed significant differences between the means of each size group (Table 1). This small single-date sample shows that chironomids of all sizes are available as food for fish and supports the inference that the maximum size of ingested chironomids is due to a factor(s) related to the fish itself and not to prey size availability. Cladocerans were the second-most numerous prey, comprising 10.2% of items measured. Thorax diameter size varied by a factor of 3.1, from 0.161–0.507 mm, considerably less than the 12.5-fold range of chironomid head widths. Unlike the chironomids, there was no obvious visual break point in the cladoceran data, but ANOVA results revealed a significant difference among the means (F(2, 27) = 9.81; P < 0.001) and Tukey-Kramer tests found a significant difference between the cladocerans ingested by the fish in group 2 (8.0–9.6 mm SL) when compared to groups 3 and 4 (group-1 fish consumed only six cladocerans and were excluded from the analysis). The difference between groups 3 and 4 was not significant (Table 1). Osteology Developing bones that are involved in food capture and manipulation were scored based on their degree of ossification in 25 specimens from 6.2–16.2 mm SL (Fig. 2). We predicted that ossification of bones over small size ranges around 9.6 mm and around 14.1 mm would permit larger food items to be ingested and processed. Ceratobranchial 5, the pharyngeal arch, which is involved in prey mastication, is partially ossified in the smallest cleared and stained specimen (6.2 mm SL) and is fully ossified around 8.8 mm SL (Fig. 2). Other key bones involved in prey capture—including the premaxilla, maxilla, and dentary— show early ossification around 7 mm SL, but they are not fully ossified until ≈10 mm SL. Around 9.6 mm SL, most bones of the feeding apparatus were nearly or fully ossified, except for cartilage-replacement bones such as the Table 1. Comparison by larval group of the mean (± SE) of the largest 10% of chironomids consumed by each group during the entire study and in a single day collection (21 Jun 2005), and a comparison of the mean (± SE) of the largest 50% of cladocerans consumed by each group during the study. Mean values in the same column with different superscript letters are significantly different from each other using a Tukey-Kramer test. Largest 10% Largest 10% chironomids Largest 50% chironomids (single day cladocerans Larval group (entire study) collection) (entire study) Group 1 0.145A (± 0.002) NA NA (6.5–7.9 mm SL) (n = 9) Group 2 0.193 B (± 0.011) 0.202 A (± 0.022) 0.225 A (± 0.009) (8.0–9.5 mm SL) (n = 10) (n = 4) (n = 4) Group 3 0.312 C (± 0.008) 0.335 B (± 0.017) 0.328 B (± 0.018) (9.6–14.0 mm SL) (n = 26) (n = 9) (n = 14) Group 4 0.439 D (± 0.031) 0.566 C (± 0.013) 0.324 B (± 0.015) (>14.0 mm SL) (n = 18) (n = 4) (n = 12) 356 Northeastern Naturalist Vol. 15, No. 3 Figure 2. Degree of ossification of bones of the feeding apparatus and fin rays in 20 Creek Chub larvae ranging from 6.2 mm to 16.2 mm SL. Dotted lines at 8.0 mm SL, 9.6 mm SL, and 14.1 mm SL indicate break points where larger prey items can be ingested. palatine, gill arches, ceratohyals, and metapteryoids. The nearly complete ossification of key food-acquisition bones by the 9.6 mm SL break point 2008 S.A. Ward and M.M. Coburn 357 suggests that ossification may be required before mesolarval Creek Chubs are able to capture chironomids with head widths greater than 0.23 mm. Fin development was also examined as a potentially important factor (Fig. 2). Caudal fin rays appear around 7.1 mm SL, and are fully ossified at 9.5 mm SL. The anal, dorsal, and pectoral fins are the next to form and their fin rays begin to appear from 9.1–9.9 mm SL, with fin ray ossification completed around 11.25 mm SL in the dorsal and anal fins. The complete ossification of fin rays in the pectoral and pelvic fins does not occur until much later in development, at about 15.5 mm SL. Thus, the best correlated features at 9.6 mm SL appear to be the ossification of bones directly associated with feeding, the ossification of the caudal fin rays, and perhaps the formation and early ossification of the anal, dorsal, and pectoral fins. At the second break of 14.1 mm SL, there is no correlation with ossification of cranial elements, nor with ossification of fin rays. Discussion In most ecosystems, fish species switch their diet as they undergo development from larval to adult stages. In general, stream dwelling larvae ingest algae and, later in development, switch to the main food item of adults and juveniles (Gerking 1994, Sanderson and Kupferberg 1999). In the present study of an urbanized stream, chironomids, not algae, were the main prey item found in the smallest Creek Chub larvae and remained the dominant prey item through development to the juvenile stage. The size of ingested prey often increases as larval fish grow and a simple ecosystem with a homogeneous food base such as Doan Brook affords the opportunity to examine size preferences within the same prey item. An analysis of chironomid head widths provided strong statistical support that the breaks in maximum prey size were, in fact, real boundaries. For Creek Chubs <8.0 mm SL, maximum chironomid head width agrees with or, in the smallest larvae, exceeds gape predictions (Fig. 1), and gape appears to be the principal limiting factor constraining prey size. From 8.0–9.6 mm SL, maximum chironomid head widths stabilized at about 0.23 mm and gape does not appear to be a limiting factor in prey selection as fish grow. At 9.6 mm SL, maximum prey size increased approximately to 0.39 mm and remained relatively constant until fishes reached 14.1 mm SL, where maximum prey size again increased with some fishes consuming chironomids with heads as wide as 0.60 mm. Cladocerans were the second most numerous prey for Creek Chubs in Doan Brook, but far fewer individuals were found in their gut contents. There was a significant break in maximum size of cladocerans ingested by fish near 9.6 mm SL, but no difference was found among cladocerans ingested by fish larger than 9.6 mm SL. Differences among preferred size of cladocerans may have been obscured by their relatively narrow size range of about 3x, whereas chironomids varied by a factor of 12.5, and it can be inferred that cladocerans are not as useful as chironomids in examining the relationship between maximum prey 358 Northeastern Naturalist Vol. 15, No. 3 size and larval fish size. Even so, they also support the hypothesis that a nonlinear increase in prey size occurs near 9.6 mm SL. The body size of Creek Chubs at which feeding is gape-limited is similar to that reported in other gape-limitation studies including Perca flavescens (Mitchill) (Yellow Perch) (<10 mm total length [TL]; Bremigan et al. 2003) and Pomoxis annularis Rafinesque (White Crappie) (≈10 mm TL; DeVries et al. 1998). Our study is unusual in that we found gape limitation occurred twice, once in the mid-mesolarval stage for fishes smaller than 8.0 mm SL, and again in the late mesolarval stage at 9.6 mm SL (Fig. 1). When they passed 14.1 mm SL, Creek Chubs ingested still larger prey, but this increase was unrelated to gape. The stepwise results of our study are not inconsistent with data from studies where individual prey size has been reported. For example, Schael et al. (1991) showed maximum prey size to be gape-limited at 10 mm TL in Yellow Perch, to remain essentially constant between 10 mm and 14 mm TL, but then to increase suddenly at 14 mm TL. Bremigan et al. (2003) graphed similar stepwise increases in the maximum size of food items consumed by larval Yellow Perch in Green Bay, WI, with plateaus between 6–8 mm TL, 8–11 mm TL, 11–14 mm TL, and >14 mm TL. Schael at al. (1991) also recorded a constant size of the largest zooplankton consumed by Pomoxis nigromaculatus (Lesueur in Cuvier and Valenciennes( (Black Crappie) between 7–11 mm TL with a doubling of maximum size at 11 mm TL. DeVries et al. (1998) graphed what may be a similar stepwise increase in both maximum and mean zooplankton size consumed by White Crappie in Clark Lake, OH. Among cyprinids, a non-linear break in prey size was found at 23 mm SL in Rutilus rutilus (L.) (Roach), which suggested an intermediate stage before the juvenile stage, in which the greatest quantity and maximum diversity of prey was ingested (Reyes-Marchant et al. 1992). However, other results demonstrate a more linear increase in maximum prey size. Aplodinotus grunniens Rafinesque (Freshwater Drum) tended to ingest prey as large as their gape permits (Schael et al. 1991), and laboratory-raised Lepomis macrochirus Rafinesque (Bluegill) regularly ingested prey larger than predicted by gape measurements (Bremigan and Stein 1994). In addition to increasing gape, the ability to ingest larger prey could be due to many factors, such as ossification of cranial elements, fin development, or greater functionality of the sensory system. Kawakami and Tachihara (2005) found that landlocked Plecoglossus altivelis ryukyuensis Nishida (Ryukyu-ayu) exhibited a diet shift, and fed upon larger prey items with increased SL. They proposed that the diet shift was coupled with increased feeding activity that was the result of increased swimming ability, enlargement of the mouth, and the development of the sense organs. Over the course of their development, Roach undergo a shift from a diet of phytoplankton to zooplankton and benthic macroinvertebrates, and this shift has been linked to intestine development (Mark et al. 1989, Reyes-Marchant et al. 1992), greater protrusibility of the mouth, and fin development (Reyes-Marchant et al. 1992), as well as body 2008 S.A. Ward and M.M. Coburn 359 shape and feeding apparatus morphology (Hjelm et al. 2003). There was little evidence in early juvenile Sciaenops ocellatus L. (Red Drum) that their prey size is constrained by gape alone, but rather, prey capture is influenced by development of other features of the feeding mechanism, such as ossification of the hyoid and opercular series (Krebs and Turingan 2003). In Creek Chubs, ossification of bones used in food acquisition is completed near the critical size of 9.6 mm SL, which suggests that complete or nearly complete ossification of these bones may be required before Creek Chubs are able to utilize larger prey items. Many species have been found to feed on prey smaller than gape limitations would suggest. For example, DeVries et al. (1998) found larval Dorosoma cepedianum (Lesueur) (Gizzard Shad) behaved contrary to expectations and continued to consume small prey even though they were no longer gape limited. Gizzard Shad have an extended larval period, and their cranial skeleton does not ossify until they attain sizes of about 20 mm SL (M.M. Coburn, pers. observ.). The relatively late ossification of the Gizzard Shad feeding apparatus may be a factor in limiting it to smaller than expected prey. The ossification sequence of cranial bones in Creek Chubs agrees well with the sequence found in Danio rerio Hamilton (Zebrafish), with ceratobranchial 5 ossifying first, followed by the opercle, and then, over a very narrow size range, ossification of the branchiostegal rays, hyomandibula, maxilla, dentary, premaxilla, ceratohyals, and other bones involved in feeding (Cubbage and Mabee 1996). Mabee et al. (2000) compared the relative sequence of ossification of cranial bones across four evolutionarily divergent fish species: Zebrafish, Barbus barbus L. (Barb), Betta splendens Regan (Bettas), and Oryzias latipes (Temminck and Schlegel) (Ricefish). They found that bones involved in feeding ossified before other bones, and that the high level of concordance in ossification sequence across species suggested that functional demands tightly constrain potential variation in this pattern. It may be that the ability to acquire larger prey items is correlated with the ossification of these bones as a general phenomenom across many species of fishes. Fin development is another factor that has been suggested to play a major role in prey capture (Reyes-Marchant et al. 1992, Webb and Weihs 1986). Changes in food habits of larval fishes in later developmental stages could be potentially linked to the complete development of fins, which would allow for more effective searching and capture. Near the break point at 9.6 mm SL, only the caudal fin of Creek Chubs is fully formed and ossified, with the anal, dorsal, and pectoral fins only beginning to form (Fig. 2). At the next break point (14.1 mm SL), there is nearly complete formation but not ossification of all fins, which may lead to increased mobility and aid in capturing larger chironomids. The correlation, however, appears to be weak, and other factors such as greater development of the sensory systems should be investigated. 360 Northeastern Naturalist Vol. 15, No. 3 In sum, developmental processes strongly influence prey selection by young fish (Mark et al. 1989). For smaller larval fishes, prey selection can be particularly confined by the morphological constraints of gape size, but the relationship between gape and prey size has been shown to vary across fish species (Bremigan and Stein 1994, Schael et al. 1991). After gape-size limitations are exceeded, many species continue to feed on smaller prey than predicted. The results of this study show that maximum chironomid head width increases in a stepwise fashion from 0.15 mm in larve <8.0 mm SL, to 0.23 mm in larvae from 8.0–9.6 mm SL, to 0.39 mm in larvae from 9.6–14.1 mm SL, and to 0.61 mm in larvae >14.1 mm SL. Gape is limiting twice during the mesolarval stage, once for fish <8.0 mm SL and again at 9.6 mm SL; the third increase in maximum prey size is not associated with the removal of a gape limitation. The correlation between non-linear increases in prey size and the ossification of the feeding apparatus and caudal fin suggests they may play an important role in prey size selection at about 9.6 mm SL. There is less evidence that ossification of the skeletal system plays a role in the shift in prey size around 14.1 mm SL, and other factors should be investigated to account for this shift. The stepwise increase in maximum prey size observed in this study could be consistent with data from studies of feeding by larval fishes on zooplankton. The much larger range of chironomid sizes as compared to zooplankton may make such shifts easier to detect. Acknowledgments The authors thank John Carroll University for its support of S.A. Ward in her Master’s research, and R. Drenovsky for assistance with statistical analyses. The Olympus IX71 inverted microscope and software were purchased with funds from NSF 0431285, Systematics of Cypriniformes, Earth’s Most Diverse Clade of Freshwater Fishes, awarded to M.M. Coburn. Literature Cited Barber, W.E., and W.L. Minckley. 1971. Summer Foods of the Cyprinid Fish Semotilus atromaculatus. Transactions of the American Fisheries Society 100:283– 289. Bremigan, M.T., and R.A. Stein. 1994. Gape-dependent larval foraging and zooplankton size: Implications for fish recruitment across systems. Canadian Journal of Fisheries and Aquatic Sciences 51:913–922. Bremigan, M.T., J.M. Dettmers, and A.L. Mahan. 2003. Zooplankton selectivity by larval Yellow Perch in Green Bay, Lake Michigan. Journal of Great Lakes Research 29:501–510. Copes. F.C. 1978. Ecology of the Creek Chub. University of Wisconsin, Museum of Natural History, Stevens Point,WI. Report on the Fauna and Flora of Wisconsin. No 12. Cubbage, C.C., and P.M. Mabee. 1996. Development of the cranium and paired fins in the Zebrafish Danio rerio (Ostariophysi, Cyprinidae). Journal of Morphology 229:121–160. 2008 S.A. Ward and M.M. Coburn 361 Cunha, I., and M. Planas. 1999. Optimal prey size for early turbot larvae (Scophthalmus maximus L.) based on mouth and ingested prey size. Aquaculture 175:103– 110. DeVries, D.R., M.T. Bremigan, and R.A. Stein. 1998. Prey selection by larval fishes as influenced by available zooplankton and gape limitation. Transactions of the American Fisheries Society 127:1040–1050. Dinsmore, J.J. 1962. Life history of the S. atromaculatus, with emphasis on growth. Proceedings of the Iowa Academy of Sciences 69:296–301. Gerking, S.D. 1994. Feeding Ecology of Fish. Academic Press, San Diego, CA, 416 pp. Hjelm, J., G.H. van de Weerd, and F.A. Sibbing. 2003. Functional link between foraging performance, functional morphology, and diet shift in Roach (Rutilus rutilus). Canadian Journal of Fisheries and Aquatic Sciences 60:700–709. Kawakami, T., and K. Tachihara. 2005. Diet shift of larval and juvenile landlocked Ryukyu-ayu, Plecoglossus altivelis ryukyuensis in the Fukuji Reservoir, Okinawa Island. Japanese Fisheries Sciences 71:1003–1009. Krebs, J.M., and R.G. Turingan. 2003. Intraspecific variation in gape-prey size relationships and feeding success during early ontogeny in Red Drum, Sciaenops ocellatus. Environmental Biology of Fishes 66:75–84. Lammens, E.H.H.R., and W. Hoogenboezem. 1991. Diets and feeding behaviour. Pp. 353–376, In I.J. Winfield and J.S. Nelson (Eds). Cyprinid Fishes: Systematics, Biology and Exploitation. Chapman and Hall, London, UK. 668 pp. Mabee, P.M., E. Aldridge, E. Warren, and K. Helenurm. 1998. Effect of clearing and staining on fish length. Copeia 1998:346–353. Mabee, P.M., K.L. Olmstead, and C.C. Cubbage. 2000. An experimental study of intraspecific variation, developmental timing, and heterochrony in fishes. Evolution 54:2091–2106. Magnan, P., and G.J. FitzGerald. 1984. Ontogenetic changes in diel activity, food habits, and spatial distribution of juvenile and adult Semotilus atromaculatus. Environmental Biology of Fishes 11:301–307. Makrakis, M.C., K. Nakatani, A. Bailetzki, P.V. Sanches, G. Baumgartner, and L.C. Gomes. 2005. Ontogenetic shifts in digestive tract morphology and diet of fish larvae of the Itaipu Reservior, Brazil. Environmental Biology of Fishes 72:99–107. Mark, W., W. Weiser, and C. Hohenauer. 1989. Interactions between developmental processes, growth, and food selection in the larvae and juveniles of Rutilus rutilus (L.) (Cyprinidae). Oecologia 78:330–337. McMahon, T.E. 1982. Habitat suitability index models: S. atromaculatus. US Department of the Interior, Fish Wildlife Service, Washington, DC. FWS/OBS- 82/10.4 23 pp. Northeast Ohio Regional Sewer District. 1992. Greater Cleveland Area Environmental Water Quality Assessment 1989–1990. Cleveland, OH. 508 pp. Potthoff, T. 1984. Clearing and staining techniques. Pp. 35–37, In H.G. Moser, W.J. Richards, D.M. Cohen, M.P. Fahay, A.W. Kendall, Jr., and S.L. Richardson (Eds.). Ontogeny and Systematics of Fishes. Special Publication 1, American Society of Ichthyologists and Herpetologists. Allen Press, Lawrence, KS. Qin, J.G., and T. Hillier. 2000. Live food and feeding ecology of larval Snapper (Pagrus auratus). Pp. 63–68, In D. McKinnon, M. Rimmer, and S. Kolkovski (Eds.). Hatchery Feeds: Proceedings of a Workshop held in Cairns. The Fisheries Research Development Corporation, Cairns, Australia. 362 Northeastern Naturalist Vol. 15, No. 3 Quist, M.C., M.R. Bower, and W.A. Hubert. 2006. Summer food habits and trophic overlap of Roundtail Chub and Creek Chub in Muddy Creek, Wyoming. Southwestern Naturalist 51:22–27. Reyes-Marchant, P., A. Cravinho, and N. Lair.1992. Food and feeding behaviour of Roach (Rutilus rutilus Linne 1758) juveniles in relation to morphological change. Journal of Applied Ichthyology 8:77–89. Rosati, T.C., J.R. Johansen, and M.M. Coburn. 2003. Cyprinid fishes as samplers of benthic diatom communities in freshwater streams of varying water quality. Canadian Journal of Fisheries and Aquatic Sciences 60:117–125. Sanderson, S.L., and S.J. Kupferberg. 1999. Development and evolution of aquatic larval feeding mechanisms. Pp. 301–377, In B.K. Hall and M.H.Wake (Eds.). The Origin and Evolution of Larval Forms. Academic Press, San Diego, CA. 448 pp. Schael, D.M., L.G. Rudstam, and J.R. Post. 1991. Gape limitation and prey selection in larval Yellow Perch (Perca flavescens), Freshwater Drum (Aplodinotus grunniens), and Black Crappie (Pomoxis nigromaculatus). Canadian Journal of Fisheries and Aquatic Sciences 48:1919–1925. Sellman, S.M., J.R. Johansen, and M.M. Coburn. 2002. Using fish to sample diatom composition in streams: Are gut contents representative of natural substrates? Pp. 529–536, In A. Economou-Amili (Ed.).16th International Diatom Symposium. Athens and Aegean Islands Proceedings. 2001. University of Athens, Athens, Greece. Snyder, D.E., M.B.M. Snyder, and S.C. Douglas. 1977. Identification of the Golden Shiner, Notemigonus crysoleucas, Spotfin Shiner, Notropis spilopterus, and Fathead Minnow, Pimephales promelas, larvae. Journal of the Fisheries Research Board of Canada 34:1397–1409. SPSS, Inc. 2003. SPSS 12.0.1 for Windows. SPSS Inc, Chicago IL Webb, P.W., and D.Weihs. 1986. Functional locomotor morphology of early life-history stages of fishes. Transactions of the American Fisheries Society 115:115–127.