Morphometry and burst swimming in six continental fish species from Costa Rica

Introduction: Current research of fish locomotion is focused on creating better underwater vehicles and how environmental stress factors modify swimming. Objective: To study the relation of morphometric characteristics with burst swimming in six representative species of continental fishes from Costa Rica. Methods: We measured total length, standard length, height and area of the tail of 38 individuals from six species. We used a Kruskall-Wallis test and a Boxplot graphic to compare species; and a PCA test to identify body variables that influence swimming. A Non-Metric Dimensional Scaling (NMDS) test was done for species and position in the water column. Results: The fastest swimming corresponded to A. nigrofasciata (9,29cm/s), while S. salvini (1,65cm/s) was the slowest. Burst swimming speed is influenced by body size and tail type, and differed with position in the water column, being surface species the fastest. Conclusions: Morphological and ecological characteristics determine differences in burst swim.

There are several classifications of swim movement in fish. According to the form of locomotion, there are two main types of locomotion: (I) undulatory movement of body and caudal fin and (II) median movement of paired fins (Blake, 1983;Videler, 1993). In regards to the temporal scale: (I) periodic or sustained swim, typically used to swim long distances at a continuous speed and (II) transient movements, such as rapid escape manoeuvres or fast movements to catch prey (Sfakiotakis, Lane, & Davies, 1999). Finally, there is also an order according to the performance: (I) Sustained, that is defined as filled aerobically for long times (II) Prolonged, that is defined as fully aerobically but lasts a shorter time period than the sustained and (III) Burst anaerobic swim that reaches maximum speed (Reidy, Kerr, & Nelson, 2000;Plaut, 2001;Robolledo et al., 2014).
The integration of how the physical environment affects the swimming biomechanics is a complex issue that incorporates ethological aspects, intra and inter specific species relations that go beyond simple body and metabolic costs (Altringham & Ellerby, 1999;Liao, 2007;Binning, Roche, & Layton, 2013;Marras et al., 2015). Environment forces have influenced the aquatic movement, incompressibility and high density which are key physical properties of water as a locomotion medium (Altringham & Ellerby, 1999). Other physical variables such as levels of oxygen, temperature, acidification and pollutants can dramatically affect the capability of locomotion on fishes (Johansen & Jones, 2011;Domenici et al., 2013;Afshan et al., 2014).
Current research of fish locomotion is focused on movements to develop autonomous underwater vehicles (Shang, Wang, Tan, & Cheng, 2012;Flammang, Tangorra, Mignano, & Lauder, 2017) and environmental stress factors that modify the swim (Johansen & Jones, 2011;Binning et al., 2013;Domenici, Herbert, Lefrançois, Steffensen, & McKenzie, 2013;Afshan et al., 2014). A significant anthropogenic stress factor is the construction of barriers such as dams and bridges (Gibson, Haedrich, & Wernerheim, 2005;Katopodis, 2005). Usually, these human structures do not consider the capacity of fish to cross them (Gibson et al., 2005;Katopodis, 2005), which can cause habitat fragmentation and population isolation (Perkin et al., 2015). Thus, it is important to contemplate the main aspect of fish locomotion for the development and design of passage, exclusion and guidance systems. The burst swim is useful to escape from predation, feeding and cross (Reidy et al., 2000;Robolledo et al., 2014) and can be correlated with the ability to survive mortality (Nelson & Claireaux, 2002).
Research into the impacts of river infrastructure on Costa Rica is related to other topics, such as mammals (Picado, Parallada, Mora, & Sánchez, 2017), plants (Benavides, Barboza, Rodríguez, & Gairaud, 2015;Campos, 2017) and habitat fragmentation (Farah, 2016). The results of this study into the morphology and ecology of a fish species in relation to the burst speed can serve as a baseline of information for the country. The objective of this study was to determine the relationship between the burst swim speed and the morphological and ecological characteristics from six representative species of riverine ecosystems in Costa Rica.

MATERIALS AND METHODS
Two field visits were made to the South Pacific and the Caribbean slope. On the Pacific; the fishes were collected from Ceibo River (9˚10'41''W -83˚20'41''N), a tributary of the Grande de Térraba. On the Caribbean; in a tributary stream of the San José River (83°52'58,34"W -10°13'12,54"N). Additional organisms raised in the laboratory were used from the Humid Laboratory of the School of Biological Sciences of the Universidad Nacional in Heredia. Species selection was based on the number of organisms collected and that were representative and common species on both slopes (Sibaja, Bussing, Garita-Alvarado, & López, 2013).
The different individuals used for the morphometric and swimming measurements were collected through field and laboratory visits. Adult and juvenile fishes were sampled with a Smith-Root Model LR-24 backpack electro fisher and purse seine. All the collected organisms were classified up to the species level, according to Bussing (2002) and the current taxonomic validity was verified (Eschmeyer, 2018).
Morphometric measurement: Data collected included total length (TL), standard length (ST), extended tail height (H) according to Holden and Raitt (1974) using a Vernier calliper. For the tail area (A) measurement, each fish was photographed and we used the UTHSCSA Image Tool for Windows 3 image analysis software to determine the area of the tail of each organism tail picture.
Burst swim estimation: The burst speed was measured in a plastic channel of a meter and a half length, width of 30cm and a height of 15cm. The channel was sealed at the ends, furthermore with a scale at the bottom and filled at 10cm of water at 20°C. After five minutes of acclimatization, the individual was subjected to a mechanic stimulus with the hand simulating an attack to induce the explosive swim (from the moment of start to the end of the movement). The velocity (v) was measured with chronometer and distance displaced in straight line . Each fish was tested in three replicas, allowing the individual to rest for one minute between each replica. The speed for each individual was standardized by dividing velocity according to their standard body length.
The averages and standard deviations of all morphological measurements (TL, SL, H, A and V) were calculated for each species (Table 1). A Kruskall-Wallis test and a Boxplot graphic determined if there were significant differences between burst swim at species level. A Principal Components Analysis (PCA) was performed to determine which body variables (Total length, Standard length, Tail height and Tail area) influence explosive swimming. Additionally, a Non-Metric Dimensional Scaling (NMDS) test was performed to show the relationship between species according to the variables previously used and the position in the water column. The statistical analyses and graphing were carried out on the software Past 3 and R 3.5.1.

RESULTS
The species with the highest rate of explosive swimming was A. nigrofasciata (9,29cm / s), while S. salvini (1,65cm/s) was the species with the lowest explosive swimming speed (Table 1; Fig. 1). A significant difference was found between the values of explosive swimming among all the species used during the present study (Kruskal-Wallis, test p<0,05, H = 16,88). The boxplot of burst speed (cm/s) per specie showed a well-portrayed difference (Fig. 1), A. nigrofasciata was the fastest swimmer and the specie who presented more variation.
The PCA matrix of morphological variables had two axes with an eigenvalue >1, which explained roughly 99,82% of the variance in the data. The PCA showed that the size of the fish and the area of the tail are the variables that influence the speed of the burst swim (  The Non-parametric Multidimensional Analysis (NMDS) test was grouped according to the speed, total length, tail area, body height and position on the water column (PWC) (Fig. 3). The ordination test differed in grouping for fishes from bottom and medium water column. The surface group overlaps with the rest of groups, being surface species the fastest.

DISCUSSION
This is the first study to describe the explosive swimming of six species of freshwater fish for Costa Rica. Previous studies have focused on migratory species of economic importance or marine species, which is why we have focused on smaller species that are less researched, with less knowledge of their ecology. Whilst this study analyzed a small portion of the species richness, the relationships between morphologic diversity among the dominant families has clear importance for freshwater ecosystems.
All the species used in this study have rounded, truncated and convex tails (Table 1). In the case of the species A. nigrofasciata, it presented the fastest explosive swim due to its body compressed laterally and truncated tail. Studies have shown that the tail of fish play an important role on the swimming process, as this body part under goes complex deformations independent of the body (curvature, area and angle) (Videler, 1993;Lauder, 2000;Lauder, Drucker, Nauen, & Wilga, 2003). Other characteristics that must be highlighted are the ecological features, such as behaviour and feeding. For example, A. nigrofasciata requires rapid movements to ram its prey (Ferry, Paig-Tran, & Gibb, 2015), while S. salvini has a slower swim speed when scraping the bottom (Barbee, 2005).
Previous studies mention that species have morphological and physiological adaptations to live in different habitats offered by the river (Barbee, 2005;Pichler & Schiemer, 2008;Kipanyula & Maina, 2016). The characteristics of the riverine ecosystem affects mobility and locomotion of species, thus changes in the ecosystems will cascade through the systems and impact all the species (Lupandin, 2005;Katopodis, 2005;Mueller, Pander, & Geist, 2011;Binning et al., 2013;Domenici et al., 2013;Afshan et al., 2014;Winemiller et al., 2016). Thus, under this premise, it is believed that S. salvini displays the slowest swim burst because it lives on the riverbed (where stones, trunks, branches, roots and other obstacles are found) so an explosive swim must be quick to escape predators, but at the same time allow it to manoeuvre to avoid obstacles.
The study suggests that morphological characters distinguishing the elongated body, truncated and convex tails display a swim faster than those fish with rounded tails ( Table 2). The body type is a fundamental characteristic that determines the explosive swim (Wardle & He, 1988). Other important factors are body phenotypes and absence of predators (Langerhans, Layman, Shokrollahi, & DeWitt, 2004;Langerhans & Reznick, 2010) which will modify the explosive swim.
The study of fish assemblages is associated with patterns of variation within freshwater ecosystems (Mims & Olden, 2012;Pease, Gonzalez-Diaz, Rodiles-Hernandez, & Winemiller, 2012). Our results suggest there are two well-portrayed assemblages regarding the position in the water column. The NMDS (Fig. 2) showed a difference between medium and bottom, while the surface species overlap with all the other groups. Based on the results, it is possible to verify that there is a similarity of ecomorphological forms among sampled species. Both lotic the morphological features as the ecological will determine differences in the explosive swimming speed.
Further research on the patterns of niche partitioning should be complemented with trophic ecology. Other important topic to be considered for future research is the study of how other physical variables modify this explosive swing (temperature, oxygen level). This will be particularly important since the country has a considerable hydroelectric dams development and such infrastructure generates a constraint for the passage of fish, changes in temperature and modification on the river flow (Arias & Alvarado, 2013;McManamay et al., 2015;Farah, 2016;Winemiller et al., 2016). We concluded that the burst swim depends of Morphological and ecological characteristics. In this way, the species A. nigrofasciata obtained the fastest burst swim to the type of tail, size of the organism and the position in the water column that it usually inhabits.

ACKNOWLEDGMENTS
Authors want to thank Helena Molina from the Biology School of the Universidad de Costa Rica, Claire Tanner from University of Bath and to our peer reviewers all their comments and advice helped to improve substantially this manuscript. Additionally, to Juan Ulloa of the Universidad Nacional, his contribution and facilities helped carry this experiment. Finally, to the Postgraduate System of the Universidad de Costa Rica (SEP).