Swimming and jumping ability of 10 Great Plains fish species

Abstract

Stream fragmentation can be detrimental to lotic fish species by preventing important life history movements. The maximum swimming speed and jumping ability of 10 stream fish species were evaluated to describe potential water velocity and height barriers to fish movement. A 10-L swim tunnel was used to test maximum swimming performance and an artificial waterfall with an adjustable weir was used to test jumping performance. All tested fish were between 30 and 100 mm total length. Mean maximum swimming velocity ranged from 37.5 ± 1.2 cm/s for mosquitofish to 65.0 ± 1.7 cm/s for largemouth bass and jumping ability ranged from 0 cm for bluegill to 13 cm for green sunfish. Differences in swimming and jumping ability demonstrate how limits to dispersal can be taxa specific and therefore, impact the conservation of rare species and management of exotic species. The values determined in this study can be useful in creating models to predict barriers to fish passage.

Keywords:

• culverts
• swimming ability
• jumping ability
• fish passage
• barriers
• prairie stream

Introduction

Stream fishes often require large-scale movements to maintain life history requirements including overwintering, spawning, and predator avoidance (Scheurer et al. 2003). Fish often have home ranges but will move to complete activities such as spawning by moving to more suitable habitat (Schlosser 1991). In prairie systems subjected to seasonal precipitation, populations are often regionally extirpated and would require individuals from connected streams to recolonize after precipitation events (Dodds et al. 2004). However, anthropogenic activities have altered the landscape and changed characteristics of streams (Matthews 1988; Dodds et al. 2004). Ninety-eight percent of land cover in the Great Plains has been altered by humans, most often through conversion to agriculture, and consequently dams and road crossing structures have been widely developed to provide access (Sampson & Knopf 1994). Dams and culverts can fragment a stream system and hinder fish and aquatic macroinvertebrates’ ability to disperse (Matthews 1988; Warren & Pardew 1998; Dodds et al. 2004; Guenther & Spacie 2006). Culverts change stream morphology by channelizing water, which can increase water velocity as it passes through the culvert. The increased water velocities can create waterfalls on the downstream end of the culvert as a result of substrate scouring. Thus, culverts can inhibit the movement of fish upstream as currents may be too swift or the waterfall created may be too high (Ficke et al. 2011).

Native fish assemblages also face challenges from the introduction, movement, and distribution of non-native fish. Non-native fish have entered most North American streams and have reduced the diversity and abundance of fish through a number of mechanisms (Moyle et al. 1986; Tyus & Saunders 2000). For example, invasive species can prey on native species and their young causing a decline in abundance (Tyus & Saunders 2000; Mills et al. 2004; Schumann 2012). Exotic species can also increase competition for resources, spread diseases, and disrupt other ecosystem functions (Moyle et al. 1986; Tyus & Saunders 2000; Mills et al. 2004). The intentional use of barriers is being considered to stop or slow the spread of non-native fish because of the potential to limit movement and distribution (Verrill & Berry 1995; Thompson & Rahel 1998).

Knowing the swimming and jumping ability of fish can be important in determining what may act as a barrier to fish passage. Models such as Fish-Xing have been created that can help predict which obstacles may be a barrier to fish passage. However, these models require accurate species-specific swimming and jumping values. In this paper, we tested the maximum swimming and jumping ability of 10 species of fish that are common in Great Plains streams in the United States.

Methods

Fish collection

Plains topminnow (lotic and lentic populations), Fundulus sciadicus, northern plains killifish, Fundulus kansae, western mosquitofish, Gambusia affinis, sand shiner, Notropis stramineus, red shiner, Cyprinella lutrensis, channel catfish, Ictalurus punctatus, black bullhead, Ameiurus melas, bluegill, Lepomis macrochirus, green sunfish, Lepomis cyanellus, and largemouth bass, Micropterus salmoides were collected from streams using seines, trap nets, and dip nets or from a broodstock pond located in Central Nebraska (Schumann et al. 2012) for testing of maximum swimming and jumping performance. Lentic plains topminnow were collected from a broodstock pond located at Sac Wilcox WMA and lotic plains topminnow were collected from a stream at Rock Creek fish hatchery in Nebraska. The size of the swim tunnel allowed for fish between 30 and 100 mm in total length to be tested. Collected fish were placed in a 378 L aerated tank for transport back to the University of Nebraska-Kearney Aquatic Laboratory and placed in a 1900 L holding tank. The water in the holding tank, swimming trials, and jumping trials was 17.5 °C. While in the holding tank, fish were fed ad libitum a mixture of frozen blood worms and frozen brine shrimp daily. Largemouth bass were fed small bluegill (25–40 mm). A water pump was placed in the holding tank to provide a light current of 5–10 cm/sec (Ficke et al. 2011). The fish were kept on the local photo period (40.6994°N) during the laboratory period.

Swimming assessment

To test the swimming ability of each species, a 10-L swim tunnel (modelled from a Loligo swim tunnel and constructed by Ace Irrigation, Kearney, NE) was used that was filled with water from the holding tank. A single fish was randomly netted from fish available in the holding tank and placed in the swim chamber. The fish was allowed 5 minutes to acclimate to the flow direction and to recover from handling. An acclimation speed of 0.5 body lengths per second was used (Ficke et al. 2011). After the acclimation period, a constant acceleration test, with an increase rate of 5 cm/sec every 10 seconds, was used to find the maximum swimming speed of each individual (Leavy & Bonner 2009). This increase was continued until the fish fatigued and was pinned against the back screen for four seconds (Leavy & Bonner 2009). Thirty individuals of each species were used for testing mean maximum swimming ability. Each individual was used once to avoid effects caused by the trial (Farlinger & Beamish 1978). If an individual did not swim, it was removed and counted as a non-performer and a new fish was used. Mean maximum swimming speed was compared between lotic and lentic plains topminnow populations with a Students t-test and significance set at α = 0.05.

Jumping ability assessment

To test jumping ability, an artificial waterfall (Figure 1) was created based on Kondratieff and Myrick (2005). The apparatus was 60 cm wide by 120 cm long by 120 cm tall. A weir was placed in the middle creating two chambers, an upper and a lower chamber that were 60 cm × 60 cm × 120 cm with a plunge pool depth of 23 cm. A water pump was used to move water from the lower chamber into the upper chamber at a rate of 72 L/min (Ficke et al. 2011). The flow of the water coming over the waterfall was approximately 42 cm/s. Ten fish of one species were randomly selected from available fish in the holding tank and placed into the bottom chamber. Trial fish were left for 24 hours or until a successful jump was observed. A video camera was used to ensure fish attempted to jump and all successful jumps were recorded. After the trial was completed, all fish were removed and the weir height was raised one centimeter. Ten new fish were then placed in the lower chamber for the next trial. If no fish reached the upper chamber, another trial at the same height was conducted. If no successful jump occurred during the second trial, this height was considered the maximum jumping performance. Only lotic topminnow were tested for jumping due to limited availability of lentic topminnow.

Figure 1. Jumping apparatus used to test jumping ability of fish.

Results

The mean maximum swimming velocity varied among the 10 species tested. Largemouth bass had the highest mean swimming velocity at 65.0 ± 1.6 cm/s and western mosquitofish had the lowest mean swimming velocity at 37.5 ± 1.2 cm/s (Table 1). Red shiner had the single overall highest maximum swimming velocity at 90.1 cm/s (Table 1). Mosquitofish had the lowest maximum swimming velocity at 50.1 cm/s (Table 1). Mean swimming velocity did not significantly differ between lotic (40.6 ± 1.13 cm/s) and lentic (38.7 ± 0.9 cm/s) plains topminnow populations (t = 1.36, df = 58, p = 0.09). All species had a positive correlation between length and velocity (Table 1) with cyprinids showing the greatest slope. The r² values for the increase in mean maximum swimming velocity and length of individual fish was variable with western mosquitofish having the lowest correlation at 0.21 and red shiners having the highest with 0.74 (Table 1).

Table 1. The mean maximum swimming velocity (±SE) and mean length (±SE) of each species of fish tested. Maximum velocity recorded for each species is also listed. The r² squared value is listed for the correlation between length of fish and maximum velocity for each species. The highest jump completed (cm) is listed for each species.

Species	Mean total length	Mean swimming velocity (cm/s)	Maximum swimming velocity (cm/s)	r^2 Value	Predictive equation	Jump height completed (cm)
Lotic topminnow	55.3 ± 1.1	40.7 ± 1.1	51.6	0.43	0.4383x + 16.42	4
Lentic topminnow	47.1 ± 1.0	38.7 ± 0.9	50.0	0.39	0.5327x + 13.601	NA
Killifish	69.3 ± 1.4	51.8 ± 1.7	73.2	0.38	0.7803x − 2.3083	6
Mosquitofish	48.4 ± 1.5	37.5 ± 1.2	50.1	0.21	0.3553x + 20.309	3
Red shiner	56.1 ± 1.8	57.3 ± 2.6	90.1	0.74	1.2792x − 14.52	4
Sand shiner	58.1 ± 1.3	62.0 ± 1.9	85.6	0.57	1.0787x − 0.7235	3
Black bullhead	59.7 ± 1.9	43.4 ± 2.1	69.7	0.46	0.7766x − 2.9674	1
Channel catfish	75.7 ± 2.1	61.8 ± 1.7	>82.1*	0.28	0.4423x + 28.328	6
Bluegill	47.0 ± 1.5	45.7 ± 1.5	63.7	0.62	0.7618x + 9.8998	0
Green sunfish	63.4 ± 2.0	49.2 ± 1.9	73.8	0.54	0.7094x + 4.2148	13
Largemouth bass	69.0 ± 1.0	65.0 ± 1.7	85.0	0.32	0.9613x − 1.3248	7

*Limitation of equipment at time fish were available.

The jumping results also varied among species, from 0 to 13 cm (Table 1). The highest jump and lowest jump were seen from the same family, with both recorded by Centrarchidae species (13 cm for green sunfish and 0 cm for bluegill).

Discussion

The results from these experiments add to the data available on mean maximum swimming velocity for Great Plains stream fish. Four of the species tested have no reported information on maximum swimming capabilities. In most cases, where species had been tested, the individuals captured in this study from the Great Plains region performed in a similar manner as previous studies from other regions. For example, the reported mean maximum swimming velocities of western mosquitofish, sand shiner, and green sunfish were within 7% of our results (Ward et al. 2003; Leavy & Bonner 2009). The red shiner tested from Great Plains populations displayed 53% lower mean maximum swimming velocities than those tested by Ward et al. (2003); however, the average length of their fish was nearly 13 mm longer than those tested in our study. Leavy and Bonner (2009) reported red shiner to have 41% higher mean maximum swimming velocity than found in our study, but reported standard length so it is not known how these sizes compared to fish from our trials. Red shiners displayed a high correlative relationship between mean maximum swimming velocity and length in our study (r² = 0.74).

The total length of fish tested appears to be related to individual performance of mean maximum swimming velocity. Within our results, it was evident that morphologically similar body shapes (i.e., members of the same family) would improve their mean maximum swimming velocity as they got larger. For example, the mean length of lentic plains topminnow (47.1 mm), lotic plains topminnow (55.3 mm), and northern plains killifish (69.3 mm) resulted in mean maximum swimming speeds of 38.7, 40.7, and 51.8 cm/s, respectively. Similar relationships were observed for the two Cyprinidae and Ictaluridae species and three Centrarchidae species tested. Thus, mean maximum swimming velocity may be in part associated with morphological body shapes similar among related species. While this relationship should be pursued in greater detail, it is evident that length of specimens tested should be reported and a range of lengths should be investigated when assessing mean maximum swimming velocities.

During periods of low stream flow, culvert hang height instead of water velocity is likely limiting fish passage. Therefore, species-specific jumping ability would dictate which perched culverts would be considered fish barriers. Species-specific differences in jumping ability described in this study may help make predictions on the effects of stream fragmentation on a species. For example, high jumping ability of species such as largemouth bass, green sunfish, and channel catfish would suggest these species may be less susceptible to fragmentation than species with low or zero jumping ability such as bluegill or plains topminnow. Indeed, fragmentation has been documented to alter fish communities and fish species that had high jumping ability such as green sunfish were less affected than others by fragmentation (Perkin & Gido 2012).

Understanding of maximum jumping ability was enhanced by the results from this study. Previous experiments quantifying jumping ability have focused on salmon and trout species (Kondratieff & Myrick 2006; Mueller et al. 2008). The only published literature available on a species that we tested was for green sunfish in which Ellis (1974) reported jumping abilities that were up to 10x the individual's body length. While these results indicate greater jumping ability than individuals we tested, the experiments conducted by Ellis (1974) included variable flow rates that might have contributed to improved jumping abilities among tested fish. Ficke et al. (2011) reported jumping ability for species present in Great Plains streams that would be similar to the results of the species we tested. Temperature and plunge pool depth were found to influence jumping abilities of individual species (Ficke et al. 2011), so it is important to be cautious in using results from any specific trials. Discretion should also be used in correlating jumping abilities of related species as members from the same family (bluegill and green sunfish) demonstrated the lowest and highest abilities in our trials.

Residence in lotic and lentic habitats may manifest into morphological changes that could alter swimming and jumping abilities. The Hawaiian goby Sicyopterus stimpsoni displayed functional trade-offs in body shape to match the prevailing environmental demand of avoiding predators or climbing waterfalls that was associated with their island stream (Blob et al. 2010). Morphological differences have been documented between lotic and lentic populations based on specific ecological parameters such as water velocity. Water velocity was found to have altered the height to length ratio of the carapace of the turtle, Pseudemys concinna (Rivera 2008) and also changed growth rates and body condition of freshwater drum Aplodinotus grunniens (Rypel et al. 2006). Morphological differences were found in blacktail shiner Cyprinella venusta as lentic individuals had deeper bodies and a shorter dorsal fin base that was located more anterior than lotic individuals (Haas et al. 2010). An effort to identify differences in swimming and jumping ability within a species that reside in variable water velocities has not been conducted. Although our results showed no significant differences between lotic and lentic topminnow, it should be noted that topminnow are backwater specialists and seek low flow areas even within a lotic system (Schumann et al. 2014). Ideally, another candidate species could be assessed to investigate variability in swimming or jumping performance associated with this microhabitat selection.

Understanding the swimming and jumping capabilities of fish has multiple applications. The swimming and jumping ability data can be used in models such as the Fish Xing Program to predict barriers in a field setting. For example, all of the sizes and species of fish tested in this study are capable of passing through a road crossing with flows of 40 cm/s or less. The data collected in this study can be used in models to construct road crossings to reduce the impact on fish movements within the stream. In some projects, various obstacles could be intentionally created to form a barrier preventing the spread of non-native species (Rahel 2013). Because the acceptable error is different in each model scenario, caution should be used when employing swimming and jumping abilities into different fish movement models. If fish movement through a barrier is critical, then specific experiments should be conducted with the use of local conditions (i.e., species assemblage, plunge pool depth, and local temperature regimes). As a point to assist any attempts to conduct jumping trials, we found the length of trials could be shortened because in all but catfish experiments, fish would try to jump within the first two hours and if not successful would stop these attempts.

Acknowledgement

This work was funded and made possible by Nebraska Department of Roads (grant M334). We thank the University of Nebraska at Kearney Biology Department for additional funding and laboratory space and equipment. We thank the Nebraska Game and Parks Commission for equipment, use of land, trucks and for the collection of hatchery fish. Ace Irrigation built the swim tunnel and we also thank Ashley Ficke for advice on using and building the jumping apparatus. A special thanks to Dave Schumacher for valuable advice on experimental design and review. In addition, we thank Dave Schumann, Mike Cavallaro, Zach Woiak, Rebecca Pawlak, Josh Kreitman, Brett Roberg, Sean Farrier, Adrienne Conley, Elisabeth Jorde, Kelly Willemssens, Guilherme Guarnieri, and Kaue Shera for assistance with fish collection and experimentation. Rick Grantham and Bruce Noden provided helpful comments on an earlier version of this manuscript. This research was approved and partially supported from the state and federal resources provided to the Oklahoma Agricultural Experiment Station.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

Nebraska Department of Roads [grant number M334].