Nesting microhabitat comparison of Central stoneroller and Bluehead chub: potential inference for host-switching by nest associates

Abstract

Several nest-building North American minnows (Cyprinidae) function as reproductive hosts to nest associates–species that require nests of other species for spawning. Understanding the microhabitat preferences of hosts can yield insight into the reproductive ecology of many species, especially of nest associates that can utilize nests of two or more hosts. We observed nests of Central stoneroller Campostoma anomalum in which several associate species were actively spawning. Bluehead chubs Nocomis leptocephalus began constructing nests two days later in the same stream, at which time associates abandoned stoneroller nests and continued on to spawn on chub nests. This presented a unique opportunity for accomplishing two objectives: (1) quantifying stoneroller nesting microhabitat preference and (2) comparing stoneroller and chub habitat preference to gain insight into the mechanisms that may drive host switching by nest associates. We measured substrate size, current velocity, water depth, and egg depth on seven paired stoneroller and chub nests, and compared these measurements to paired microhabitat measurements at a randomly selected point near each nest. Repeated measures analysis of variance with post hoc Tukey tests revealed that stonerollers exhibited distinct nesting microhabitat preferences from chubs. Gravel on stoneroller nests was considerably smaller than on chub nests and stonerollers nested in shallower depths than chubs. However, both species nested at similar current velocities. If nest associates switch partners based on the physical characteristics of nests, then substrate size is likely the most important factor. The larger gravel sizes on chub nests likely provide better egg aeration than stoneroller nests. Chub nests may also be safer for associate broods because male Bluehead chubs cover eggs with gravel after spawning; stonerollers do not. Future work should take an experimental approach to elucidate these mechanisms.

Keywords:

• nest association
• spawning
• microhabitat
• Nocomis
• Campostoma

Introduction

Minnows (Cyprinidae) represent one of the most diverse and widespread freshwater fish families on earth (Howes 1991). Despite their ubiquity, many contemporary studies of cyprinid autecology are limited to imperiled or geographically rare species (e.g., Gibson et al. 2004; Billman et al. 2008) or introduced/invasive species (Schrank et al. 2001; Weyl et al. 2009). Many life history studies of common, native cyprinids were conducted in the first half of the twentieth century. While these studies have been very informative, they are mainly descriptive. The large amount of descriptive studies, coupled with their commonness, has led to the notion that many common cyprinids are thoroughly studied, and has directed modern ecologists to focus on more ‘complex’ topics. However, a lack of quantitative life history information can serve as a barrier to more complex studies that require this information. These barriers are particularly evident for common cyprinids, as their ubiquity and abundance can cause them to function as keystone species in aquatic ecosystems (Power et al. 1985; Vives 1990).

Central stoneroller Campostoma anomalum (hereafter, ‘stoneroller’) is one of the most common minnow species in upland streams of the eastern United States, ranging from the Atlantic coast to New Mexico, and from northern Wyoming to north-eastern Mexico (Jenkins & Burkhead 1994). Because of its broad distribution and geographic separation of populations, C. anomalum is not monotypic; at least six distinct evolutionary entities are currently recognized (Blum et al. 2008). Several studies have described stoneroller spawning behavior (Smith 1935; Miller 1962, 1964). In general, stonerollers spawn in mid-spring in pit-like nests dug out by large males. Some stonerollers may also spawn on unmodified substrate in riffles (Miller 1962), on gravel mound nests constructed by male Nocomis chubs (Reighard 1943), or even on redds of rainbow trout Oncorhynchus mykiss (Lennon & Parker 1960). Although stoneroller spawning behavior is well observed, no studies have quantified preference for nesting microhabitat variables such as substrate size, depth, and current velocity.

The lack of quantitative nesting microhabitat information for stoneroller is particularly unfortunate because this species often functions as host to reproductive nest associates, fishes (typically other cyprinids) that utilize nests built by other species for spawning (Johnston & Page 1992). Throughout its range, many nest associate species utilize stoneroller nests for spawning. However, stoneroller seldom functions as the only reproductive host for a given nest associate species. Other nest building minnows such as Nocomis chubs, Creek chub Semotilus atromaculatus, Fallfish Semotilus corporalis, and Cutlips minnow Exoglossum maxillingua also construct nests that attract breeding associates. Further, most nest associates do not exhibit fidelity to a single host. Many associates are capable of opportunistically switching hosts (Pendleton et al. 2012), even when multiple host species are spawning in relative spatiotemporal proximity (Grady & Cashner 1988).

Although alternate host use by a range of nest associates is fairly well documented (see Jenkins & Burkhead 1994 and references therein), the mechanisms that cause associates to switch hosts remain uninvestigated. Nest building fishes can facilitate associate reproduction through the unique physical characteristics of their nests and by providing parental care to broods (Vives 1990; Johnston 1994a). Accordingly, whether or not a species exhibits nest associative spawning (or lack thereof) can have important implications for persistence in changing environmental conditions (Hitt & Roberts 2011; Peoples & Frimpong 2013). Because these facilitative mechanisms are likely to differ among hosts and associates, a better understanding of the microhabitat characteristics of nests of various hosts is a logical starting point for understanding host switching in this system.

During daily minnow nest surveys in May 2014 in a small tributary to the New River, VA, we observed several stoneroller nests being actively excavated by adult males and being used for spawning by several nest associates including Mountain redbelly dace Chrosomus oreas, Rosyside dace Clinostomus funduloides, and Rosefin shiner Lythrurus ardens (Figure 1). After two days of active spawning, Bluehead chubs Nocomis leptocephalus (hereafter, ‘chubs’) began construction of their large gravel spawning mounds (see Maurakis et al. 1991 and Sabaj et al. 2000). At the onset of chub spawning, all associates immediately abandoned stoneroller nests and began actively spawning on chub nests. Clearly, associates gained some form of fitness advantage to utilizing nests of both host species for spawning. These proximate behaviors provided a unique opportunity to accomplish two objectives: (1) quantify spawning microhabitat preferences of Central stoneroller and (2) compare the microhabitat characteristics of stoneroller and chub nests. We hypothesized that, behavioral differences of hosts and associates aside, the physical characteristics (i.e., substrate sizes, water depth at placement, and current velocity) of stoneroller and chub nests would differ, thus providing insight into the reasons for host switching by nest associates.

Figure 1. A tuberculate male Central stoneroller (center) tends his nest while Rosyside dace and Mountain redbelly dace congregate for spawning opportunities.

Methods

We conducted this study in an approximately 400-m reach of Toms Creek in Montgomery County, VA (N 37.261955, E -80.436715). Toms Creek is a third Strahler-order tributary to the New River (Ohio/Mississippi drainage) in the Valley and Ridge physiographic province of the central Appalachian Mountains. This reach of Toms Creek was relatively narrow (mean width is 3.5 m) and shallow (maximum depth of 0.53 m). The channel was heavily entrenched, and incised banks regularly sloughed into the channel. Accordingly, substrates consisted primarily of sand, heavily embedded gravel, and outcrops of hardpan clay.

We surveyed for stoneroller and chub nests daily from 12 to 26 May 2014. Flow conditions during this time were uncharacteristically stable; no major rain events elevated the water levels and all surveys were conducted at ‘baseflow’ conditions (although no flow gaging station is located on this stream). Two workers searched for nests by walking on both banks several times each day, wearing polarized sunglasses to reduce glare. We are confident that few nests went overlooked because (a) nests of both species are very conspicuous and attract spawning aggregations of brightly colored fishes, (b) the stream is quite small, and (c) because we searched for nests thoroughly and systematically. We conducted nest observations according to Peoples et al. (2015). Briefly, we observed each nest for at least half an hour during each day of active spawning to identify species associated with the nests. To minimize the chance of not detecting a species on a nest, we made overhead videos of nests (approximately 20 minutes of video per nest) and revisited most nests multiple times each day. We reviewed videos in their entirety to ensure all species present on the nests were identified.

To assess stoneroller spawning microhabitat selectivity, we compared microhabitat variables at stoneroller nests to paired, randomly chosen points approximately 5 m upstream from each nest (sensu Wisenden et al. 2009; Bolton et al. 2015). Hereafter, we refer to these measurements as ‘unused.’ To better understand why nest associates abruptly switched from stoneroller nests to chub nests, we also compared microhabitat variables on each stoneroller nest to paired chub nests within 10 m of the stoneroller nest within the same pool. A nest was not included in the analysis if it did not occur in the same mesohabitat (pool, riffle or run; Frissell et al. 1986) as a nest constructed by the other species. We measured microhabitat variables on each nest as soon as spawning was finished, which was between two and four days after nest construction. First, we measured nest dimensions. We then measured intermediate diameters (Bain & Stevenson 1999) of 25 randomly selected substrate particles from each nest; these were compared to 25 paired randomly selected substrate particles of unused, available substrate (Wisenden et al. 2009). Next, we measured current velocity at nests using a Rickley Model 6200 current meter at the point on nests where fishes were depositing eggs. For stonerollers, observations revealed that this point was directly in the center of the nest. For chubs, this was in spawning troughs constructed by the tending male chub at the upstream, center of the nest (Maurakis 1998). We then measured water depths in which nests were placed; these measurements were made directly upstream of nests, just before the point at which nests began. This avoided bias associated with nest height.

We used repeated measures analysis of variance (ANOVA) to test for global significance in water depth, current velocity, and substrate diameters among the three microhabitat measurement locations (stoneroller nests, chub nests, and unused locations) using the lmer function in the lme4 package of R version 3.0.3. Repeated measures ANOVA is equivalent to a paired t-test with multiple factors, and accounts for the fact that each paired set of observations is nested within the identity of the location (stoneroller nest, chub nest, or unused location). We then used a post hoc Tukey test to conduct pairwise means comparisons among factors that were adjusted for experiment-wise Type-I error. All analyses were conducted at α = 0.05.

Results

We observed and measured seven paired stoneroller and chub nests. Only seven stoneroller nests were encountered in the study area, one in each of the seven pools. Paired chub nests were constructed in nearby pool tails; only the first chub nest adjacent to a stoneroller nest was used, although at least two nests were constructed in each pool by the end of the reproductive season. Accordingly, no stoneroller nest was observed that was not adjacent to a chub nest. All seven stoneroller nests were first observed on the morning of 22 May 2014; all nests had presumably been constructed on the night of 21 May and early hours of 22 May. Water temperature was 17.4 ˚C at the onset of stoneroller spawning. Two of the seven stoneroller nests were swarmed with actively spawning Rosyside dace, Mountain redbelly dace, and Rosefin shiner (Figure 1). The remaining five stoneroller nests were being used only by stonerollers. By the morning of 24 May, chubs had begun constructing nests and all nest associates abandoned stoneroller nests and began actively spawning on chub nests. Several smaller tuberculate male stonerollers also frequented chub nests, exhibiting potential spawning, feeding, and/or disruptive behaviors described in Sabaj et al. (2000). We also observed White shiner Luxilus albeolus and Crescent shiner L. cerasinus spawning on chub nests, but not stoneroller nests (Peoples et al., 2015). No additional species were observed in videos that were not observed by workers in the field.

Repeated measures ANOVA revealed significant differences among species and unused measurements for substrate size (F_2,6 = 67.4, p < 0.0001), current velocity (F_2,6 = 3.9, p = 0.0485), and water depth (F_2,6 = 7.1, p = 0093). Stoneroller nests exhibited unique physical characteristics, compared to unused locations. Substrate sizes on stoneroller nests were 2.4 mm larger than unused locations (Z₆ = 3.8, p = 0.0004). Much of the unused substrate consisted of fine gravel and coarse sand; male stonerollers excavated this substrate to reveal the larger gravels beneath. Stoneroller nests were also constructed in microhabitats that were 8.1 cm shallower (Z₆ = 3.5, p = 0.0484) than unused locations, and in current velocities 0.2 cm/s slower than unused locations (Z₆ = −7.8, p = 0.0147; Table 1).

Table 1. Spawning microhabitat variables measured at seven nests of Central stoneroller and Bluehead chub, compared to paired samples in Toms Creek, Montgomery County, VA, USA in May 2014.

	Stoneroller	Chub	Paired
Substrate diameter (cm)	9.2 ± 0.51	14.1 ± 0.51	6.8 ± 0.32
Current velocity (cm/s)	0.50 ± 0.03	0.60 ± 0.04	0.70 ± 0.11
Water depth (cm)	22.4 ± 4.6	35.3 ± 2.9	30.6 ± 2.9

Stoneroller nests were physically different from chub nests. First, chub nests are gravel mounds that are elevated over the natural, unmodified substrate, while stoneroller nests are pits in which finer substrates have been removed to reveal the larger substrates beneath. Stoneroller nests averaged 35.1 ± 1.9 cm in length, 31.1 ± 3.0 cm in width, and were 6.9 ± 1.0 cm deep, while chub nests averaged 43.7 ± 4.9 cm long, 43.7 ± 3.9 cm wide, and were 9.3 ± 1.7 cm tall. Chubs chose gravel that averaged 4.9 mm larger than gravel found on stoneroller nests (Z₆ = −11.4, p < 0.0001). Stonerollers typically nested at the margins of pool tails, while chubs usually built nests in the thalweg of pool tails, which was most often in the center of the channel. Accordingly, water depths at stoneroller nests averaged 12.9 cm shallower than chub nests (Z₆ = −3.7, p < 0.0001), although both species were selected for similar current velocities (Z₆ = −1.6, p = 0.2162). Mean substrate diameters of chub nests were more than twice as large as those in paired samples (Z₆ = −11.4, p = <0.0001) but water depths and current velocities at chub nests did not differ from unused locations (Z₆ = −1.4, p = 0.3600, and Z₆ = 1.1, p = 0.5047, respectively; Table 1).

Discussion

Understanding the spawning habitat requirements of reproductive host fishes can yield interesting insights into reproductive interactions. This study presents the first quantitative description of central stoneroller spawning microhabitat preference, although some qualitative descriptions of habitat and behaviors have been presented such as communal broadcast spawning in pool tails (Smith 1935; Lennon & Parker 1960; Miller 1962; Miller 1964; Jenkins & Burkhead 1994). Microhabitat variables measured in this study match the general qualitative descriptions provided by Miller (1962) and Lennon & Parker (1960). Central stoneroller is widespread throughout the eastern United States and Canada, and at least six distinct evolutionary entities of this polytypic species are recognized. In fact, Blum et al. (2008) suggest that populations of C. anomalum in the upper New River basin represent a yet-to-be-described species (Campostoma sp., the Virginia stoneroller). Similar cladistic patterns are evident in Nocomis, including the potential elevation of a new species endemic to the upper Roanoke and New River drainages (Nocomis sp. cf. leptocephalus, Nagle & Simons 2012). Quantitative nesting microhabitat descriptions of several Nocomis species have been presented (Lobb & Orth 1988; Maurakis 1998; Wisenden et al. 2009; Peoples et al. 2014), including for the Roanoke/New River form of Bluehead chub in the present study (Bolton et al. 2015). A more thorough understanding of Campostoma nesting habitat requirements across a larger geographic extent will yield better insight into the reproductive ecology of these species, and by extension their nest associates. This information may be particularly useful as the true diversity of these genera continues to be revealed (Blum et al. 2008; Cashner et al. 2010; Nagle & Simons 2012).

Nesting microhabitat preference of stonerollers differed from chubs. First, substrate sizes on chub nests were considerably larger than on stoneroller nests. The mechanism behind this pattern is obvious; chubs have large, subterminal mouths that can carry larger stones than stonerollers, which have smaller, inferior mouths that are suitable for manipulating only small stones or for digging in finer substrate. Larger substrate is important for facilitating aeration of eggs (Maurakis et al. 1992). Furthermore, stonerollers constructed nests in the margins of pool tails, while chubs typically placed nests in the thalweg or area of most flow. Pool tails are often locations of downwelling where oxygen-rich water flows downward through nests (Baxter & Hauer 2000). By placing nests outside of this area, stonerollers may not be taking advantage of this hydraulic characteristic, despite the fact that they selected for current velocities similar to chubs. Chub nests can be highly permeable (Wisenden et al. 2009), sometimes just as permeable as the water column itself (Peoples et al. 2014). While no study has examined permeability of stoneroller nests, it is possible that their construction of smaller, subsurface gravel in pool margins (rather than in areas of direct flow) allow for less egg aeration than chub nests.

Larger substrate sizes may also be related to safety from predation. Nesting male Bluehead chubs may attack the occasional marauding egg predator (often tuburculate male stonerollers), but do not guard broods through larval development like sunfishes (Centrarchidae) (Jenkins & Burkhead 1994). Instead, chub and associate eggs must be protected mainly by the coarse gravel in which they are deposited. Large interstitial spaces between gravel on chub nests allow eggs to sink several centimeters through the nest (Maurakis et al. 1992). Furthermore, tending male Bluehead chubs continue to deposit gravel on nests, effectively burying eggs well within the nest (Maurakis et al. 1991). Stoneroller nests, with their finer gravel and pitted structure, probably have smaller interstitial spaces for eggs, possibly making them more exposed to predation.

Stonerollers also nested at shallower depths than chubs. This difference is unlikely to be the reason that associates switched from stoneroller to chub nests. In general, shallower habitats should provide safety from large-bodied predators of both eggs (namely Northern hogsucker Hypentelium nigricans and White sucker Catostomus commersoni), and adults (namely Rock bass Ambloplites rupestris) (sensu Power et al. 1985). In which case, stoneroller nests would be safer than chub nests, and nest associates would not have switched hosts. Furthermore, egg depth did not differ between nests of the two species. Although it is possible that egg depth measurement may have been biased on chub nests because eggs can sink into interstitial spaces, a potentially deeper egg depth would not make chub nests more favorable for the previously stated reasons. Both these observations suggest that nest depth probably plays a diminished role in determining which nests are chosen by associates.

Stonerollers and chubs selected similar current velocities for nesting; this is obviously not the reason nest associates switched hosts. In choosing pool margins, stonerollers selected for distinctly lower-than-unused flows. On the other hand, chubs exhibited no flow specificity in selecting nesting habitat. Nocomis are generally highly selective for current velocity at nesting sites (Lobb & Orth 1988; Maurakis 1998; Wisenden et al. 2009). However, Bolton et al. (2015) noted that in small streams, Bluehead chub nested in slower-than-average microhabitats during years of high flows, but exhibited little current velocity preference during spawning seasons with low instream flow such as in 2014.

Temperature may be another mechanism driving host choice by associates. Reported initial spawning temperature of Central stoneroller is considerably lower than Bluehead chub (13 ˚C versus 15 ˚C in Virginia, Jenkins & Burkhead 1994). Further, a weekly study of cyprinid reproductive condition revealed that in 2013, Central stonerollers began spawning several weeks prior to Blueheadchub in a tributary to the James River, Virginia (S. Floyd, B. Peoples and E. Frimpong, unpublished). Reported initial spawning temperatures of C. oreas and C. funduloides in southwestern Virginia are nearer to chubs than stonerollers (Jenkins & Burkhead 1994). Accordingly, most large, nest-capable stonerollers may be post-nuptial by the time many nest associates reach reproductive condition. However, in 2014, no stoneroller reproduction was observed until two days prior to the onset of chub reproduction. It is possible that the particularly harsh winter and cold spring of 2013–2014 delayed stoneroller reproduction in Toms Creek to a point closer to when associates had reached reproductive condition. This may explain why associates readily began spawning in stoneroller nests, but switched to chub nests as soon as they were constructed. Temperature-driven spawning cues between hosts and associates should be examined in systems in which stoneroller or other species serves as the primary reproductive host (Starnes & Starnes 1981; Mattingly & Black 2013; White & Orth 2014).

Nest association is a widespread and complex reproductive behavior among fishes and can be an important mechanism structuring stream fish communities (Walser et al. 2000; Hitt & Roberts 2011; Pendleton et al. 2012; Peoples & Frimpong 2013). However, we still know relatively little about the basic mechanisms that determine host and associate behavior (Wallin 1992; Johnston 1994a; Shao 1997b) or interaction outcomes (Fletcher 1993; Johnston 1994b; Shao 1997a). For instance, a logical next step would be to compare fitness outcomes of associates on nests of both host species. Furthermore, understanding the consequences of nest association in a single-host environment would also yield insight into the mechanisms that drive nest association. This study is clearly limited by spatiotemporal extent; information among years and across a more diverse range of abiotic contexts will surely contribute to a better understanding of the various factors leading to host choice by nest associates. Furthermore, we only consider physical characteristics of nests as a potential driver of host switching. Behavioral mechanisms may also be at work in this system, particularly the important effect of parental care (Johnston 1994a). For example, tending male Bluehead chubs cover eggs with gravel after spawning, while stonerollers abandon their broods relatively quickly (Sabaj et al. 2000). As another alternative, host switching by nest associates may represent a form of ‘bet hedging’ in which having some eggs in the nests of each host minimizes the risk of large-scale brood failure. Clearly, much more research is needed to fully understand this interaction. We intend this work to encourage future experimental approaches to understanding the mechanisms that drive host switching in this system, and for understanding the community-level implications for the differences in reproductive success that must accompany such choices.

Acknowledgements

This work was funded by NSF DEB Award #1120629 to EAF. Data on specific nest use by each associate species is available from the Dryad Digital Repository (doi:10.5061/dryad.2ds6h).

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

NSF DEB [award number 1120629].