Movement and Longevity of Imperiled Okaloosa Darters (Etheostoma okaloosae)

OKALOOSA Darters (Etheostoma okaloosae) are small percid fish that are geographically limited to six small stream systems in northwest Florida. Okaloosa Darters were originally listed as endangered in 1973 and were recently downlisted to threatened status in 2011 (USFWS, 2011). The decision to change the conservation status of this imperiled species was based in part on the availability of data describing key aspects of its biology and life history, including fecundity, spawning behavior, body size, longevity, population size and distribution, population genetic structure, and diet (Collette and Yerger, 1962; Burkhead et al., 1992; Dorazio et al., 2005; Jordan et al., 2008; Austin et al., 2011). In contrast, no published data were available on the movement and site fidelity of Okaloosa Darters. This gap is notable given that understanding species-specific home ranges, movement rates, and dispersal patterns is important to effective management and conservation of fish populations (e.g., Labbe and Fausch, 2000; Shute et al., 2005; George et al., 2009).

Okaloosa Darters inhabit stream margins where woody debris, detritus, root mats, vegetation, and fine flocculent materials form a complex habitat (Jordan et al., 2008). Individuals are rarely observed in habitats lacking this complexity, especially not in open stretches of bare sand that characterize stream channels in the region (FJ and HJ, unpubl.). During visual surveys, darters are often found in the same patches of habitat over multiple years, suggesting that individual Okaloosa Darters may not move great distances (Jordan et al., 2008). The presumably low mobility and isolation of Okaloosa Darters to small habitat patches along the margins of streams prompted a more rigorous investigation into their movement behavior.

For hundreds of years various types of marks have been used to study movement of fishes (McFarlane et al., 1990). This extensive body of research has focused primarily on larger, more robust species that could tolerate a range of different types of marks. Tracking the movement of small-sized fishes has proven more difficult because many types of marks have negative effects on survival and behavior of target species. This challenge has been overcome through the development of marks such as visible implant elastomer (VIE; Northwest Marine Technologies, Inc.), which do not appear to adversely affect survival or behavior of small, marked fishes (e.g., Roberts and Angermeier, 2004; Weston and Johnson, 2008).

Marks not only provide a means of tracking movement, but also provide valuable information concerning longevity. Estimates of age using scale and otolith techniques can show great variation between interpreters (Campana, 2001) and age estimates based on visual inspection of length-frequency histograms are somewhat subjective. For example, previous research based on analysis of length-frequency histograms suggests that Okaloosa Darters typically live two years and die after their second year of spawning (M. F. Mettee, unpubl.) or typically live three years (Burkhead et al., 1992; V. E. Ogilvie, unpubl.). In contrast, recapturing a marked fish provides an exact age from the time of initial marking onward. Accurate estimation of age and longevity is essential for modeling population dynamics (Campana, 2001).

The present mark–recapture study was designed to explore the efficacy of using VIE to mark small fishes and to evaluate movement, site fidelity, and longevity of Okaloosa Darters.

MATERIALS AND METHODS

Laboratory study

We conducted a pilot study on syntopic, non-imperiled Brown Darters (E. edwini) to evaluate the effects of VIE marking on survival and mark retention prior to marking Okaloosa Darters. Brown Darters were collected from Long and Rocky creeks (Okaloosa County, Florida) between 14 and 16 July 2004 using a 3 mm mesh seine. Brown Darters (mean 34 mm standard length [SL], range 29–41 mm SL) were transported in aerated coolers back to Loyola University New Orleans, where they were transferred to two 133 l aquaria, kept on a 12L∶12D photoperiod, and fed a daily diet of frozen brine shrimp. Fish were allowed to acclimate for one week prior to applying a VIE mark. Using a 29-gauge needle and without the administration of anesthesia, a 5–6 mm-long mark was injected approximately 1–2 mm lateral (with left or right side arbitrarily chosen) of the first dorsal fin (Fig. 1). Yellow VIE was used because it provided the greatest contrast to the natural coloration of both Brown and Okaloosa Darters. Additionally, Largemouth Bass (Micropterus salmoides) foraging in well-lit and structure-free laboratory aquaria showed no preference for Brown Darters marked with red, yellow, or clear VIE (DH and FJ, unpubl.) A total of 51 Brown Darters were marked and thereafter checked for mortality and mark retention daily for a period of 30 days. At the end of the test period, fish were euthanized in tricaine methanesulfonate (MS-222), preserved in a 10% formalin solution, and added to the Loyola University New Orleans vertebrate teaching collection.

View Figure

• Download as Powerpoint
• Download Figure

Field study

We marked all fish captured within a single 20 m reach at each of six creeks throughout the range of Okaloosa Darters: Ben's Creek, Juniper Creek, Long Creek, Point Lookout Creek, Swift Creek, and Turkey Creek. Capturing, marking, releasing, and 24 h observations of Okaloosa Darters occurred on 14–20 August 2004. We captured Okaloosa Darters from each 20 m reach in three consecutive passes using mask, snorkel, and small plastic nets (Jordan et al., 2008), which typically accounts for over 90% of the fish present in a stream reach (Dorazio et al., 2005). Sites were selected from long-term monitoring sites with stable Okaloosa Darter populations and representative habitat. Fish captured from the left and right banks were placed in separate 19 l containers and marked on the left or right side of the dorsal fin, respectively, using the same technique used to mark Brown Darters in the laboratory. Marked Okaloosa Darters were placed in 19 l aerated recovery containers, and five minutes of recovery time were allowed between the marking of the last individual and the time of release. Fish were released simultaneously at the 10 m midpoint of the stream reach along the side corresponding to their capture. A total of 275 individuals were marked across all six sites (Table 1). Due to concerns about increased mortality due to handling time, fish lengths were not measured during the marking procedure. Standard lengths of marked fish were measured annually until the final recapture period in 2012.

Table 1.  Number of Individuals Recaptured at Each Site during Each Resampling Period. Top number shows number of individuals recaptured. Middle number in parentheses indicates the percentages of the originally marked that were recaptured. The bottom number shows the number of individuals that were found on the opposite stream margin from their original capture and marked side.

View Fullscreen

We counted (Jordan et al., 2008) marked Okaloosa Darters within each 20 m reach after 24 h and one month later (24 to 26 September, 2004), and then recaptured Okaloosa Darters each August for the following eight years. When marked individuals were found, their positions were flagged in the stream and noted as having a left or right mark. The distance each individual had moved from the release point was measured after 24 h and one month, but not for annual recaptures. An extra 10 m upstream and an extra 10 m downstream of the 20 m reaches were resampled after 24 h and one month, but not for annual recaptures. We tested for a difference in the distribution of marked Okaloosa Darters within the 20 m reaches between 24 h and one month using a Kolmogorov-Smirnov test. Dispersal data for the six sites were combined for each of the two sampling periods to increase sample size. Levene's test was also performed to test for heterogeneity of variance between these two time periods, and Chi-square tests were used to determine whether a difference existed in the number of fish that moved upstream and downstream. Pearson correlations were also conducted to determine whether crossing of Okaloosa Darters from one margin to the other was correlated with stream width, stream depth, or canopy cover. All statistical tests were conducted using PASW-18.0 (SPSS Inc., 2009. PASW Statistics for Windows, Version 18.0. Chicago, IL).

During the final annual census in August 2012, we measured the SL of all Okaloosa Darters captured in 20 m reaches from four of the above streams and in two additional streams where the long-term abundance of Okaloosa Darters is being monitored. Length probability shadowgrams were constructed to estimate the number of age classes present in each of these six stream reaches. Shadowgrams are composite graphics of 11 histograms with various bin sizes superimposed using semi-transparent kernel density smoothers to facilitate interpretation of continuous distributions (SAS Institute Inc., 2012. JMP 10 Basic Analysis and Graphing, Second edition. SAS Institute Inc., Cary, NC).

RESULTS

Laboratory study

There was no mortality associated with marking Brown Darters with VIE during the 30-day course of the study. Mark retention was 100% 24 h post marking, 96% (49/51) 48 h post marking, and 94% (48/51) 72 h post marking. Retention remained at 94% through termination of the study at 30 days.

Field study

Average recapture rates of Okaloosa Darters across the six study sites were (mean±standard deviation) 31±15% 24 h after marking (marking only occurred at the onset of the study), 45±8% after one month, 22±10% after one year, and decreased continuously thereafter at most sites until no marked fish were captured. The only exception to this was Turkey Creek, where a single recapture was made at years four, five, and seven. No marked Okaloosa Darters were recaptured during the annual census in August 2012.

The distribution of Okaloosa Darters within 20 m reaches was similar between the 24 h and one-month recapture periods (two-tailed Kolmogorov-Smirnov test, Z  =  1.254, P  =  0.086; Fig. 2). Likewise, Levene's test indicated homogenous variance between the 24 h and one-month distributions (F  =  1.442, P  =  0.231). A total of 47 fish moved an average of 5.7±4.9 m upstream, and a total of 34 fish moved an average of 5.2±4.1 m downstream of the release points after 24 h. At one month, 53 fish moved an average of 6.1±4.0 m upstream, and 67 fish moved an average of 6.6±5.5 m downstream of the release points. There was not a statistically significant difference in the number of fish that moved upstream or downstream at 24 h (χ²  =  2.086, df  =  1, P  =  0.149) or at one month (χ²  =  1.633, df  =  1, P  =  0.201).

View Figure

• Download as Powerpoint
• Download Figure

Study streams were relatively shallow (0.3±0.1 m depth) and narrow (5.4±1.2 m width), with a high degree of canopy cover (90.9±13.0%) and moderate flowing water (0.7±0.1 m/s). Despite narrow stream widths, most Okaloosa Darters were recaptured from the same side of the stream where they were originally captured, marked, and released (Table 1). At sites where crossing over was observed, 15.9±13.8% of marked Okaloosa Darters had crossed to the opposite side of the stream after 24 h and 14.3±13.1% had crossed after one month. Long Creek was the only study site in which crossing over was observed beyond one month, with 50% of the individuals found on the opposite side at two years and 33% at three years. The rate at which Okaloosa Darters crossed stream channels was unrelated to stream width, stream depth, or canopy cover after 24 h and after one month (Pearson correlation analysis, P > 0.05).

Standard lengths of the single recapture at Turkey Creek at years four, five, and seven were 30, 32, and 33 mm, respectively. The length probability histogram at Turkey Creek indicated the presence of four cohorts: a juvenile cohort centered around 16 mm SL and three adult cohorts centered around 25, 30, and 36 mm SL (Fig. 3). Ben's Creek also supported four cohorts, whereas Point Lookout Creek and Juniper Creek supported six cohorts, Mill Creek supported five cohorts, and Little Rocky Creek supported two cohorts (Fig. 3). Length probability shadowgrams at all sites showed the largest cohort mean SL between 30 and 35 mm. For all sites except Little Rocky Creek, the largest cohort was bounded by two cohorts at an interval of 5.1±0.9 mm (an estimate of the adult growth rate). Juniper Creek was the only site to show an additional adult cohort at 41 mm SL (Fig. 3). Cohorts below 25 mm SL were not spaced as uniformly among the different sites as those above 25 mm SL.

View Figure

• Download as Powerpoint
• Download Figure

DISCUSSION

The high rate of survival of Brown Darters marked with VIE and maintained in the laboratory strengthens support for use of this technique to safely mark small-sized fishes, including imperiled species (Roberts et al., 2008; Phillips and Fries, 2009). Fin clips result in similarly high rates of survival for Brown Darters, but clipped fins regenerate relatively quickly and are therefore unreliable for long-term marking studies (Champagne et al., 2008; but see Weston and Johnson, 2008). In contrast, most Brown Darters retained their VIE marks during our one-month pilot study in the laboratory and other species have high retention rates of VIE marks over much longer periods (e.g., Roberts and Angermeier, 2004; Weston and Johnson, 2008; Phillips and Fries, 2009). Based on qualitative observations, we hypothesize that a small number of fish lost their marks because the injection was too shallow or because the fish was very active during the recovery period and therefore caused a portion of the uncured VIE to be ejected from the puncture wound. Incomplete closure of the puncture wound due to partial ejection of the implant may lead to complete ejection of the implant soon after marking.

Field recapture rates for darters marked with VIE vary dramatically: 10–15% in E. flabellare, 20% in E. podostemone, and 25–30% in Percina roanoke over approximately three months (Roberts, 2003); 5% in P. rex over various time periods (Roberts et al., 2008); and 34% in E. neopterum over various time periods (Noel, 2012). Our 24 h (31%), one-month (46%), and one-year (22%) recapture rates were comparatively high, likely because Okaloosa Darters are relatively sedentary, congregate in complex microhabitats along margins of small streams, and are efficiently detected and recaptured using a visual census method (Dorazio et al., 2005; Jordan et al., 2008). Given that flooding can strongly affect movement of stream fishes (Albanese et al., 2004), our high recapture rates after one-month are particularly striking because Hurricane Ivan had made landfall 114 km west of our study sites about one week earlier. Study sites were exposed to high winds and extensive flooding, and discharge in the nearby Escambia River increased twenty-fold due to the hurricane (Hagy et al., 2006).

We are not certain as to why our recapture rates were higher after one-month than after 24 h. We did not test cortisol or other stress related indicators, and therefore cannot comment on the stress levels at the two time periods. However, the behavior of fish after marking was distinctly lethargic relative to pre-marking for at least a few hours. For example, marked fish would typically sit at the bottom of the recovery bucket without moving, whereas unmarked fish were naturally more mobile. Because fish movement plays a large part in the ability of a diver to detect presence, more lethargic fish would be more difficult to detect. Although movement levels seemed to increase within about an hour of marking (the time between the first fish marked and the time of release), slightly more lethargic marked fish may have played a part in our lower count at 24 h.

Relatively high recapture rates within 10 m of the release point, along with the lack of movement directionality (movement upstream and downstream of the release point), indicate high site fidelity. The lack of a difference in distribution within the stream between the 24 h and one-month recapture periods suggests that site fidelity is persistent over time or that dispersal rates are very slow. Recapture rates within the study sites remained high compared to other darter studies even at one year, and a single individual remained in a study reach for at least seven years. These field observations are supported by analyses of the mitochondrial cytochrome b gene and nuclear microsatellite loci that indicate minimal genetic mixing among, and to a lesser degree within, the six streams inhabited by Okaloosa Darters (Austin et al., 2011). Site fidelity is generally higher for fish populations within structurally complex habitats (Albanese et al., 2004), and small-bodied fishes tend to occupy smaller home ranges (Gerking, 1959; Reed, 1968; Freeman, 1995).

We were unable to systematically search for marked Okaloosa Darters outside of our 20 m stream reaches due to logistical constraints. A few marked fish were detected during cursory searches outside of the 20 m reach on Long Creek, indicating that Okaloosa Darters emigrated out of our study area and that our study likely provides a conservative estimate of movement. The ‘restricted movement paradigm’ (Gowan et al., 1994) posits that mark–recapture studies such as ours are more likely to detect non-movements than movements and lead to downward bias in movement. Also, movement at both 24 h and one month showed a leptokurtic distribution, which is indicative of the restricted movement paradigm (Fausch et al., 2002), and suggests that although the majority of individuals show little movement, a few may move long distances. Leptokurtic distributions have been observed in other studies of small, benthic fishes that undertake long distance movements (Petty and Grossman, 2004; Roberts et al., 2008; Breen et al., 2009; Hudy and Shiflet, 2009). We believe that movement upstream and downstream of the study site was not hampered by unfavorable habitat because suitable habitat was continuous in the margins of the stream as far as we observed both upstream and downstream of the study sites. Future mark–recapture studies of Okaloosa Darters and similar small stream fishes should systematically re-census, but not recapture, over a larger distance to reduce downward bias in movement estimates.

Crossing a relatively open stream channel to move between preferred habitats is likely a daunting task for relatively small fishes and therefore such movements may depend upon the presence of a corridor. Widely recognized as an important feature in terrestrial systems, less attention has been given to the importance of corridors in lotic environments. However, movement of stream fishes between preferred habitats does appear to be facilitated by the presence of corridors (e.g., Gilliam and Fraser, 2001). Movement of darters due to environmental variation differs among species and reflects differences in spatial scale, habitat preferences, and life history traits (Roberts and Angermeier, 2007). For Okaloosa Darters, the presence of a corridor appears to be necessary to facilitate movement across stream channels—even those channels that are relatively narrow, shallow, and have a well developed canopy overhead. Okaloosa Darters are cryptic when viewed against their preferred habitat along stream margins, but are easily seen by potential predators when sitting on open sand (Fig. 1). Risk of predation causes fragmentation of stream fish populations (Gilliam and Fraser, 2001) and likely explains why Okaloosa Darters were disinclined to move across streams that lacked corridors comprised of logs and associated debris or beds of submerged aquatic vegetation such as Juncus effusus, Schoenoplectus etuberculatus, or Sparganium americanum. Although based on qualitative observations, corridors persist in some stream reaches for extensive periods as the divers inspect linking vegetation patches and woody debris year after year.

Longevity of etheostomines has been estimated using length frequency data or examination of scale and otolith annuli for a large number of species and averages 3.1±0.8 years (based on 34 species listed in de Magalhaes and Costa, 2009). Previous examinations of length frequency data suggested that Okaloosa Darters lived from two to three years (Burkhead et al., 1992; M. F. Mettee, unpubl.; V. E. Ogilvie, unpubl.). Our relatively high recapture rates at two to three years post marking, along with the number of age classes suggested by our length frequency shadowgrams, indicates that average life expectancy of Okaloosa Darters should be revised upward. Assuming that individuals ≥25 mm SL are reproductively mature (Weil et al., 2012; V. E. Ogilvie, unpubl.), we found three to four cohorts of adults and one to three cohorts of juveniles, indicating that the life expectancy of Okaloosa Darters is at least four and possibly up to eight years. If age and length relationships are estimated assuming constant growth rates, an initial size of 5.5 mm SL at hatching (Collette and Yerger, 1962), and our observed adult growth rate of 5.1±0.9 mm per year. For example, five-year-old Okaloosa Darters should range from 27 to 35 mm SL. These predictions correspond with the average SL of fish recaptured from Turkey Creek at years four (30 mm) and five (32 mm) given that Okaloosa Darters were at least one year old when originally marked.

These conclusions are based on the assumption of constant growth rates, and that cohorts are representative of individuals born within the same year. Okaloosa Darters are not known to reproduce year round, although they do have a long breeding season extending from late March through October, peaking in April (Weil et al., 2012). It is likely, therefore, that Okaloosa Darters are fractional spawners, and that young are produced over several months. The cohorts shown in our length frequency distributions may show less resolution due to variation in fish size because of an extended spawning season. Also, it is likely that growth rates are not constant in adults, and likely slow with age (as suggested by the von Bertalanffy growth model). Obviously, further work will be required to achieve the resolution necessary to determine whether these cohorts represent inter- or intra-annual age classes.

Possibly the most intriguing aspect of this study is that we recaptured an Okaloosa Darter in Turkey Creek seven years post marking and, because all fish were at least one year old at initial marking, this individual was ≥8 years old. Yolk Darters (E. juliae) previously held the record for oldest etheostomine fish at seven years based on scale analysis (Hill, 1968). Okaloosa Darters are therefore the longest lived etheostomine. Based on the age and length relationship described above, an eight-year-old Okaloosa Darter should range from 39 to 53 mm SL. The eight-year-old fish recaptured at Turkey Creek was only 33 mm SL and therefore much smaller than predicted. This further suggests that growth slows with age, and emphasizes the importance of further studies to help refine our understanding of growth rates and age classes in the Okaloosa Darter.

The geographic range of Okaloosa Darters is almost entirely contained within Eglin Air Force Base, which has undertaken significant efforts to reduce sediment erosion, replace and improve stream crossings, and restore and improve stream habitat for this and other aquatic species. These conservation efforts played a significant role in the USFWS down-listing Okaloosa Darters from endangered to threatened status (USFWS, 2011). The decision to down-list was facilitated by the availability of critical information concerning distribution and abundance (Dorazio et al., 2005; Jordan et al., 2008) and population genetic structure of Okaloosa Darters (Austin et al., 2011). The ultimate decision to remove Okaloosa Darters from the Endangered Species List will require a more complete understanding of their biology. The present study fills some critical gaps by showing that Okaloosa Darters typically move relatively short distances, have high site fidelity, do not traverse across open channels that lack habitat corridors, and can live substantially longer than previously suggested. These data are already being used to inform decisions about removal of impoundments and poorly designed stream crossings and about restoring damaged stream reaches (Jelks et al., 2011).