Female downstream-hatching migration of the river shrimp Macrobrachium ohione in the lower Mississippi River and the Atchafalaya River.
Macrobrachium ohione (Ohio shrimp) is one of six species in this freshwater genus that is found throughout coastal rivers and their estuaries in the Gulf of Mexico and the southeastern Atlantic Coast (Hedgpeth, 1949; Home and Beisser, 1977; Bowles et al., 2000). This shrimp is the only member of the genus that permanently inhabits the Mississippi River System (MRS) (Taylor, 1992; Bowles et al., 2000; Barko and Hrabik, 2004). The distribution of M. ohione in the MRS historically ranged north to the Missouri River and well into the lower Ohio River. Prior to the 1940s, commercial fishermen reported large harvests of M. ohione, as part of a fishery for bait and human consumption, from as far north as Kentucky, Missouri and Illinois (McCormick, 1934; Hedgpeth, 1949; Taylor, 1992; Bowles et al., 2000). Since then, collections of M. ohione have become rare in the upper MRS, and when shrimps were taken from the northern reaches of the river, no gravid females were observed (Taylor, 1992; Conaway and Hrabik, 1997; Poly and Wetzel, 2002; Barko and Hrabik, 2004). Macrobrachium ohione has not been studied extensively within the MRS. A better understanding of this species' life history may provide insights into the decline of M. ohione populations along its northern distribution. Such information will be invaluable in the conservation of present populations and their restoration in the upper MRS.
Abbreviated or direct larval development has allowed most atyid and many palaemonid shrimp species to adapt to exclusive freshwater living (Hubschman and Broad, 1974; Jalihal et al., 1993; Bauer, 2004). However, Macrobrachium ohione is among those that have not adapted completely to fresh water. Even though juveniles, adults and larvae (stage-1 "nonfeeding" zoeae) occur in fresh water, studies suggest that M. ohione larvae have extended marine planktonic development with several zoeal stages (Dugan, 1971). The stage-1 zoea (hatching stage) will not undergo its initial molt to the stage-2 (first feeding) zoea until it reaches saline waters (Dugan et al., 1975; Jalihal et al., 1993; Bauer and Delahoussaye, 2008). McDowall (1992) termed this life cycle "freshwater amphidromy" because most life stages and breeding are conducted in fresh water, but larval development takes place in estuarine or marine habitats. Amphidromy is frequently found in fishes, crustaceans and snails that inhabit relatively small, shallow, short streams and rivers of subtropical and tropical islands (Holmquist et al., 1998; Blanco and Scatena, 2005; McDowall, 2007; McRae, 2007; Kikkert et al., 2009). However, some species of amphidromous shrimps occur in rivers far upstream from the sea (Hartmann, 1958; Ibrahim, 1962; Ling, 1969; Magalhaes and Walker, 1988). In the MRS, M. ohione has been observed up to 1500 km upstream from the Gulf of Mexico (Conaway and Hrabik, 1997; Barko and Hrabik, 2004; Bauer and Delahoussaye, 2008). Only a few studies (Bauer and Delahoussaye, 2008; Rome et al., 2009) have addressed how these shrimps can accomplish amphidromy within a large, long river system.
Within a large fiver like the Mississippi River (MR), the mechanism of larval delivery to saline waters becomes critical for amphidromous species. In amphidromous shrimps, alternative hypotheses on larval delivery to the sea are (1) that adults hatch the larvae upstream and allow the river currents to carry the larvae out to sea or (2) that adults migrate to or near marine environments to release larvae (Bauer, 2004). Recent studies on Macrobrachium ohione from the Atchafalaya River (AR) have observed significant increases in stage-1 larvae mortality after 3 d of exposure to fresh water (Bauer and Delahoussaye, 2008). Optimal molting success from stage-1 to stage-2 zoea occurs when larval drift in fresh water is no longer than 1-3 d (Rome et al., 2009). This suggests that larval delivery to coastal estuaries by river drift alone is not likely for far upstream populations of M. ohione in the MRS.
As in other caridean shrimps, Macrobrachium ohione females incubate embryos below the abdomen until hatching (~18 d; Bauer and Delahoussaye, 2008). During the breeding season, observations of such gravid females in the coastal waters of Barataria Bay (Gunter, 1937), the lower AR (Truesdale and Mermilliod, 1979; Bauer and Delahoussaye, 2008) and Galveston Bay in Texas (Reimer et al., 1974) suggest that females may migrate downstream into coastal bays. Dingle (1996) proposed that migrations are critical to many organisms and are selected for over time for various purposes (e.g., to ensure the most favorable environmental conditions are inhabited, dispersal into new habitats, predator avoidance, to provide an adequate gene pool for future generations). Given (1) the salinity requirements for M. ohione larval development, (2) the occurrence of populations far upstream in the MRS and (3) the seasonal occurrence of reproductive females in coastal waters, it is reasonable to hypothesize that a downstream hatching migration has evolved in M. ohione to deliver larvae to estuaries and marine waters.
The objective of this study was to test the hypothesis that gravid females migrate downstream to deliver larvae to coastal estuaries in the lower MRS for larval hatching and development. Predictions from this hypothesis were tested by sampling and analyzing spatial-temporal distributions of reproductive-sized females (RSF) in the lower MR and AR in Louisiana. Sampling of both "upstream" and "downstream" sites in the MR as well as the AR allows comparison of the seasonal distributions of female Macrobrachium ohione based on reproductive condition and size. If downstream hatching migrations occur, downstream sites were predicted to have greater proportions of females with embryos at any stage of development (IE), females with near-hatching embryos (NHE) and females with near-spawning ovaries (NSO) than at upstream sites within the Mississippi and Atchafalaya deltas during the reproductive season (Apr.-Aug.). As caridean shrimps from warmer waters can produce multiple broods per reproductive season, the spawning pattern (e.g., the presence of mature ovaries while incubating embryos) of reproductive females may suggest multiple spawning events in the deltas of both rivers if a downstream hatching migration occurs. Furthermore, increases in the relative abundance of RSF at the downstream sites were predicted during the reproductive season if a mass migration of females takes place.
In Louisiana, the MR flows from north-central Louisiana in a southeasterly direction through Baton Rouge and empties into the Gulf of Mexico south of New Orleans. The AR, a distributary of the Red River and MR, flows south from its origin at the Old River Control Complex near Lettsworth, LA, and empties into the Gulf of Mexico via the Atchafalaya Bay (Fig. 1). In each river system, an upstream and downstream site was sampled (Fig. 1). The MR downstream location was Pass A Loutre (PAL) Wildlife Management Area (29[degrees]22.2'N, 89[degrees]32.2'W) within the Mississippi Delta. The upstream MR location was at the Entergy Corporation nuclear power plant, River Bend (RB) Station (30[degrees]43.4'N, 91[degrees]21.1'W), 411 km upriver of PAL near St. Francisville, LA. In the AR, samples were taken downstream at the Atchafalaya Delta (AD) Wildlife Management Area (29[degrees]41.6'N, 91[degrees]12.9'W). The upstream site in the AR was a private dock, 146 km upstream of the AD, located at Butte La Rose (BLR), LA (30[degrees]19.6'N, 91[degrees]41.7'W). Staff of the Louisiana Department of Wildlife and Fisheries (LDWF) conducted the sampling at both downstream sites.
SPATIAL-TEMPORAL DISTRIBUTION OF INCUBATING AND REPRODUCTIVE-SIZED FEMALES
Sampling began at the two delta sites in Apr. 2008, and because of very high water conditions at both upstream sites, 1 mo later at BR and 2 mo later at RB. Sampling then continued through Oct. 2009. Sampling techniques were similar to those described in Bauer and Delahoussaye (2008). Samples were collected using baited shrimp traps constructed from "hardware" cloth (6.4 mm mesh size). Traps consisted of two separate pieces: (1) a barrel measuring 76 cm X 117 cm, closed at one end with wires and clamps; (2) a funnel measuring 38 cm X 91 cm with a 2-3 cm opening serving as the entrance into the trap and connected to the open end of the barrel. Sampling events took place twice per month at all sites, and each sample consisted of catch from three traps set for two consecutive nights. The traps were baited with one perforated can of commercial cat food and set with the funnel facing downstream. Samples were initially fixed in a 10% freshwater-formalin solution prepared in the field with ambient river water. Upon returning to the laboratory, samples were rinsed of formalin with tap water and stored in 70% ethanol.
[FIGURE 1 OMITTED]
For measurements on reproductive parameters and relative abundance, random subsamples of shrimps were taken from samples containing >>>200 shrimps. First, the entire sample was emptied into a container divided into four quadrants. Then shrimps were dispersed among the quadrants by shaking the container. After being dispersed, all shrimps in a randomly chosen quadrant (by coin toss) were removed. If the initial quadrant did not contain 200 shrimps, then the last step was repeated until the count reached at least 200 shrimps.
To examine spatial differences of reproductive females (females incubating embryos) during the reproductive season, the reproductive condition of females was monitored using methods described in Bauer and Delahoussaye (2008). To determine the sex of any individual that was not obviously female (i.e., shrimps with embryos or maturing ovaries), the presence (male) or absence (female) of an appendix masculina on the endopod of the second pleopod (swimmeret) was recorded. To assess the developmental stage of a female's embryos, a dissecting microscope was used to view the brooded mass under the shrimp's abdomen. The embryonic developmental stage was ranked as 0 = no embryos present beneath the abdomen; 1 = only yolk visible (newly spawned); 2 = blastodisc present but no eye pigmentation; 3 = eye pigmentation or eye development visible, cephalothorax and pleon not separated; or 4 = egg present, little yolk, carapace and pleon separate (Bauer and Delahoussaye, 2008). For statistical analyses, embryo stages ranked 1 or 2 were considered early developing embryos (EDE) while stages 3 and 4 were considered near-hatching embryos (NHE). Ovarian maturity in females was ranked by using a dissecting microscope to view the ovaries through the transparent carapace of the shrimp. Rankings for ovarian maturation were 1 = no ovarian development observable, 2 = ovary developing but not extending into the anterior carapace space (ACS), 3 = ovary extending into and up to half of the ACS, or 4 = ovary filling more than half of the ACS. Reproductive-sized females (see below) with ovarian maturation ranked 1 or 2 were considered early developing ovaries (EDO) and those ranked 3 or 4 were considered near-spawning ovaries (NSO). All proportion data is based on samples containing [less than or equal to] 200 shrimps or subsamples summing at least 200 shrimps.
Relative abundances, measured as Catch Per Unit Effort (CPUE), were documented to determine if there was an influx in the number of RSF at the downstream sites during the reproductive season. This influx would be indicative of a mass movement of RSF to coastal waters. Reproductive-sized females were defined as any female with a body size [greater than or equal to] the smallest female with embryos across all samples in the respective river system that a female was sampled. Body size was measured as carapace length (CL), the chordal distance (mm) from the posterior edge of the eye orbit on the mid-dorsal edge of the carapace (Bauer and Delahoussaye, 2008). In the MR, a RSF was any female measuring [greater than or equal to] 7.8 mm CL. In contrast, the RSF of the AR population was [greater than or equal to] 9.6 mm CL.
Each sample event was planned to consist of three traps sampling over two nights. However, accidents of sampling (e.g., unfavorable weather, unexpectedly high or low fiver height variations, trap theft) sometimes resulted in a lost trap or night of sampling. Therefore, the CPUE of a sample was calculated as [SIGMA] # RSF/(# traps * # nights). Note that the total number of shrimps considered for CPUE was limited according to the subsample protocol described above. Only one sample or no samples were taken at (1) PAL in 2008: Apr., Aug., Oct. and Dec., 2008; 2009: Jun., Sept.; (2) RB, Oct., 2009; (3) AD, Sept., 2008; 2009: Apr., May, Jul.; and (4) BLR Sept., 2008. Samples during these periods were not taken because of very unfavorable weather or river conditions and/or lack of LDWF personnel at downstream sites.
All statistical analyses on the distribution of female shrimps based on their reproductive condition during the breeding season were conducted separately for both fiver systems (MR and AR) and years sampled (2008 and 2009). Fisher's Exact test (Proc Freq., SAS Institute Inc., 2007) was used to test the null hypothesis of no positive association between the downstream sampling site (PAL and AD respectively) and (1) females with embryos at any developmental stage, (2) females with NHE and (3) RSF with NSO. Bauer and Delahoussaye (2008) provided a priori information that allows use of this directional test (Gotelli and Ellison, 2004). Fisher's Exact Test (Proc Freq., SAS Institute Inc., 2007) was also used to test the null hypothesis of no positive association between the stage of embryonic development and the stage of ovarian maturation (successive spawning) at each site. Reproductive-sized females sampled during the reproductive season were pooled together for distribution analysis based on their reproductive condition at all sites. For all statistical tests, significance was established at [alpha] = 0.05.
The null hypothesis that the interaction between location (downstream = PAL and AD; upstream = RB and BLR) and season (reproductive season = RS, nonreproductive season = NR) produced no difference in the CPUE of RSF was tested using the Kruskal-Wallis test (Proc NPARIWAY, SAS Institute Inc., 2007). This non-parametric analysis was used because CPUE data did not meet the assumptions of the parametric A.NOVA test. The RS was defined as the months of Apr. to Aug.; the months from Sept. to Mar. were designated as the NR. Samples were then grouped by location, season and year for tests of the hypothesis of no difference in the median CPUE among groups. Statistical significance was established at [alpha]= 0.05.
In both years of the study, females were observed incubating embryos from Apr. to Aug. (reproductive season). Contrary to the migration hypothesis, in the MR the proportion of females incubating embryos at any stage of development was smaller at the downstream site (PAL) than that observed upstream at RB (Fig. 2A). In the MR, there were no positive associations between the proportion of females incubating embryos (at any stage of development) and PAL in 2008 or 2009 (Fisher's Exact test: 2008, n = 1925, P = 0.06; 2009, n = 1096, P = 1.0). However, in the AR larger proportions of incubating females were observed downstream (AD) when compared to BLR for 2008 and 2009 (Fig. 2B). The increase in the proportions of incubating females was significantly associated with the AD (Fisher's Exact test: 2008, n = 924, P [less than or equal to] 0.01; 2009, n = 793, P [less than or equal to] 0.01). Odds ratios suggest that in the AR there were 4.0 (95% CL = 3.0-5.2) and 5.6 (95% CL = 4.1-7.6) times the chances of observing a female incubating embryos downstream than upstream in 2008 and 2009 respectively.
Although the MR lacked a spatial distinction in the proportion of females incubating embryos, a closer examination of the developmental stage of incubated embryos indicated that proportions of females incubating near-hatching embryos (NHE) was greater at the downstream site than the upstream site for both years of the study (Fig. 2A). Significant associations were identified between PAL and females with NHE in 2008 and 2009 (Fisher's Exact test: 2008, n = 1084, P [less than or equal to] 0.01; 2009, n = 511, P [less than or equal to] 0.01). At PAL (MR) in 2008 and 2009, the odds of observing females with NHE were 8.9 (95% CL = 5.7-15.8) and 5.5 (95% CL = 2.3-3.5) times greater than observing females with NHE at RB. In the AR, larger proportions of females incubating NHE were observed downstream than upstream in both years of the study. However, observations in the AR from 2008 indicated no significant association (Fisher's Exact test, n = 404, P = 0.22) between females with NHE and the AD. In contrast, 2009 observations suggest that there was a significant association (Fisher's Exact test, n = 400, P = 0.01). In addition, when the two reproductive seasons are considered together, odds ratios indicate that at AD there were 1.44 (95% CL = 1.04-1.97) times the odds of observing females incubating NHE than BLR.
In the MR, the proportion of RSF with near-spawning ovarian development (NSO) was smaller at PAL in 2008 and larger in 2009 when compared to RB (Fig. 2A). The 2008 observations showed no significant association (Fisher's Exact test, n = 1925, P = 0.82) between RSF with NSO and PAL in 2008, but in 2009 the association was significant (Fisher's Exact test, n = 1096, P [less than or equal to] 0.01). Similar results were observed in the AR. The proportions of RSF with NSO were larger downstream than upstream (Fig. 2B), although the association was significant for 2008 (Fisher's Exact test, n = 924, P [less than or equal to] 0.01) and not significant in 2009 (Fisher's Exact test, n = 793, P = 0.06).
[FIGURE 2 OMITTED]
Some caridean shrimps from tropical and subtropical zones are capable of having multiple broods within a breeding season, which is indicated by observations of females with NSO while incubating NHE (Bauer, 2004). At downstream sites in both rivers, numerous reproductive females were observed with NHE and NSO suggesting that some females, after hatching a brood in the estuary, would spawn another brood (Figs. 3A, B). The positive association between NHE and NSO was significant in 2008 and 2009 PAL and AD samples (Fisher's Exact test: 2008 PAL, n = 981, P [less than or equal to] 0.001; 2009 PAL, n = 300, P [less than or equal to] 0.001; 2008 AD, n = 267, P [less than or equal to] 0.001; 2009, n = 306, P [less than or equal to] 0.001). However, a greater number of females at downstream sites with NHE had ovaries with no or little ovarian development (EDO), indicating that successive brooding is not the only reproductive pattern observed in these shrimps (Figs. 3A, B).
At upstream sites, many females carried embryos at an early stage of development (EDE) but had little ovarian maturation (EDO), indicating that they had recently spawned and were not yet prepared to hatch larvae (Figs. 3A, B). This association of females with EDE and EDO at upstream sites was significant (Fisher Exact test: 2008 RB, n = 103, P [less than or equal to] 0.001; 2009 RB, n = 211, P [less than or equal to] 0.001; 2008 BLR, n = 137, P [less than or equal to] 0.001; 2009 BLR, n = 94, P [less than or equal to] 0.001).
Reproductive-sized females were obtained at all locations except AD in Oct. 2008 and RB during winter sampling (Dec. 2008 to Feb. 2009). As predicted by the migration hypothesis, the median CPUE of RSF was greater at PAL during the reproductive season when compared to the other season/site combinations (Fig. 4A). However, CPUE was greater downstream than upstream during the nonreproductive season as well, which may not suggest an influx of RSF from upstream sites. In the MR for both years of the study, the null hypothesis of no difference in CPUE of RSF among season/site combination was rejected (Kruskal Wallis test: 2008, [[chi square].sub.3] = 18.7, P = 0.0003; 2009, [[chi square].sub.3] = 13.5, P = 0.004). However, Dunn's multiple comparison did not find any significant differences between the median CPUE of RSPAL, NRPAL and RSRB (i.e., no seasonal influx of RSF downstream) in 2008 or 2009 (Fig. 4A). Similar spatial-temporal differences were found in the CPUE of RSF in the AR for 2008 (Kruskal Wallis test, [[chi square].sub.3] = 11.9, P = 0.008), where an increase in catch was observed downstream between seasons. In that year the greatest median CPUE was observed at BLR during the RS and no significant differences were identified between RSAD, RSBLR and NRBLR (Fig. 4B). In the AR, 2009 sampling identified no significant differences in CPUE of RSF among season/site combinations (Kruskal Wallis test, [[chi square].sub.3] = 7.52, P = 0.057).
Seasonal increases in abundance of reproductive females of Macrobrachium ohione in coastal waters during the spring and summer breeding season have been hypothesized to result from a downstream hatching migration (Reimer et al., 1974; Truesdale and Mermilliod, 1979; Bauer and Delahoussaye, 2008). In this investigation, we tested this hypothesis in two major rivers (Atchafalaya and Mississippi) in southern Louisiana. Various measures of reproductive activity support the hypothesized downstream migration of females to hatch incubating embryos in the estuaries of the MR and AR. Significantly larger proportions of females with near-hatching embryos (NHE) were sampled at the coastal sites than the upstream sites. In the MR and AR, numerous females incubating NHE also have NSO, suggesting that multiple spawning events occur downstream. At upstream sites, many females carried embryos at an early developmental stage and had little ovarian development, indicating that hatching will occur further downstream, i.e., in the estuary. We also observed other females upstream with NHE and NSO; these females may hatch larvae before arriving at the estuary, but the larvae would be within drifting distance of it. Furthermore, such females will have soon spawned another brood, which will hatch at the estuary, the optimal hatching environment. These observations suggest that there is a downstream movement of reproductive shrimps in the AR and MR, which apparently has evolved as a mechanism to deliver larvae to saline coastal waters in time for continued larval development. This evidence agrees with other studies on M. ohione that reveal that larval development occurs in estuarine or marine waters (Dugan et al., 1975; Bauer and Delahoussaye, 2008) and that female shrimps migrate to coastal estuaries to delivery hatch larvae (Bauer and Delahoussaye, 2008).
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
Although statistical analyses did not provide overwhelming evidence for the influx of RSF at the delta sites of both river systems, the greatest median CPUE of RSF were found at the downstream sites during the reproductive season for all samples except for AR in 2008. The AR 2008 sample did have a temporal increase at AD and a spatial-temporal increase when comparing RSAD to NRBLR. This may or may not be indicative of a downstream migration, especially when considering that at PAL the CPUE of RSF was greater there when compared to upstream sites regardless of season. However, it does indicate that more potentially reproductive females are observed within closer proximity of the saline environments needed for larval development. This suggests that these females will have greater reproductive success than those found at farther upstream habitats. In both river systems, statistical tests may have been affected by the small sample size. A more extensive study (e.g., increase replication and a longer multiyear study) may provide sufficient data to show statistical significance.
Rome and collaborators (2009) observed that the amount of time larvae can drift in fresh water before reaching saline water is limited. That study provided further evidence for the downstream hatching migration of female Macrobrachium ohione. The stage-1 (nonfeeding) M. ohione larvae can optimally drift no longer than 3 d in fresh water before arriving in waters of sufficient salinity to stimulate molting to the stage-2 (first feeding) and thus on to full larval development (Rome et al., 2009). Using unpublished observations of river velocity (from river stage data) collected from 1975-1983 at Baton Rouge, LA, by the United States Army Corps of Engineers (USACE), we estimate that the MR mean optimal larval drifting distance was 441 km (266 km minimum, 637 km maximum) during the 2008 and 2009 reproductive seasons. These estimates suggest that RB (411 km upstream) is on average ~3 d from the MR delta site during the reproductive season (Apr.-Aug.). We also found that reproductive intensity tended to peak during periods of higher water and greater flow velocity (May and Jun.). On the days sampled during this period of our study, the river stages at RB averaged a gauge height of 13.3 m. At a river stage of 12.2 m (max river stage provided in the USACE data), the surface water velocity is 8.8 km [h.sup.-l] (USACE, unpubl.). Larvae drifting at this velocity would take ~2 d to reach PAL from RB during the peak of reproductive season. Using the observations on water velocities and drifting distances reported in Bauer and Delahoussaye (2008) for AR, BLR is 146 km upstream of the AD, which is just within the maximum optimal drifting distance (3 d) for M. ohione (Rome et al., 2009). These estimates indicate that at river stages and water velocities found during the MR and AR spring river floods, stage-1 larvae hatched at these upstream sites can reach coastal saline waters in sufficient time to stimulate further larval development. However, stage-1 larvae released farther upstream would have little or no opportunity to reach water of sufficient salinity in time to promote continued larval development.
Nevertheless, we propose that Macrobrachium ohione females do not depend on river drift alone as their only method of larval delivery to estuarine or marine habitats for the following reasons. When the data for both years of the study are combined and the number of females with NHE at upstream and downstream sites are divided by the number of gravid females at each site respectively, it was observed that at RB only 32% of gravid females were incubating NHE. In contrast, nearly 80% of incubating females at the PAL carried NHE. This implies that females in coastal waters are ready to hatch embryos near the sea. In addition, Rome et al. (2009) reported that stage-1 larvae were significantly more abundant in plankton tows conducted in coastal waters (AD) versus the BLR upstream site, suggesting that most females hatch their embryos at coastal sites. Finally, the previous estimates on optimal larval drifting distances are probably underestimated because they do not include the additional unknown distance the larvae must travel after arriving at the deltas to reach salinities necessary for molting. The MR and AR spring floods occur during the M. ohione hatching season, so that the delta areas have salinities of <1 ppt. The larvae probably have to move even farther into coastal bays or offshore to encounter the salinities of 10-15 ppt optimal for complete larval development. Without incubating females migrating closer to the coast before hatching larvae, populations farther upstream than BLR and RB would contribute little to the recruitment of new individuals into M. ohione populations.
The MR population does not fit all aspects of the AR model of female downstream migration proposed by Bauer and Delahoussaye (2008). Unlike the AR population, there was no spatial difference in the proportion of females incubating embryos at any developmental stage, and the association of the proportion of RSF with NSO at PAL varied in significance between years. This deviation from the model observed in the AR might be explained by the spatial extent of MR populations and the morphology of the MR when compared to the AR. Within the MR, populations of reproductive females have occurred from as far north as southern Missouri ~1560 km from Gulf of Mexico (Conaway and Hrabik, 1997; Bowles et al., 2000; Barko and Hrabik, 2004). At RB during the reproductive season, there may be a continuous stream of incubating females arriving from farther north on their way down to the MR delta which may explain the lack of a positive association between the proportion of incubating females and PAL. The distance the shrimps must travel to coastal bays may also affect the association between incubating females (with embryos at any developmental stage) and the sampling sites. Bauer and Delahoussaye (2008) reported that Macrobrachium ohione incubate their embryos on average for 18 d. Larger, presumably older females sampled at RB may come from far upstream in the MR, hatch larvae prior to reaching the downstream sampling site and return to upstream swimming. This would account for sampling of these incubating females at the upstream site and not the downstream site, i.e., the river estuary where larvae need to be hatched. This hypothesis may also explain why there was no significant association between ovarian condition and the downstream site in the MR. Due to the distance the shrimps must travel, females with NSO are potentially continuously observed upstream but will not be sampled downstream prior to a change in their reproductive status. In summary, incubating females from far northern populations may be making the hatching migration towards the Gulf of Mexico but simply do not make it in time for release of larvae far enough downstream to reach saline water in time for successful development.
In contrast, the northern range of the AR population terminates at the Old River Control near Lettsworth, LA, 250 km from the coast. It is feasible that the more upstream populations may be depleted of incubating females and females with NSO during the reproductive season because these females are moving downstream. Furthermore, the limited length of the AR may allow the females to conduct the entire reproductive migration to the AD; but this is difficult to confirm without a viable method of tracking individual migrating females. Even with the differences between river systems and year to year variation observed in this study, there is still strong support for the reproductive downstream migration of females.
Macrobrachium ohione's distribution within the northern MRS has been a topic of debate because the life history, at least in southern populations, is amphidromous. Although the species is now rare in the upper MRS, abundant populations once occurred in the MR as far north as 1500 km from the Gulf and up into the lower Ohio River (McCormick, 1934; Hedgpeth, 1949; Conaway and Hrabik, 1997; Bowles et al., 2000; Barko and Hrabik, 2004). Is it possible for shrimps with a life expectancy of perhaps ~1-2 y (Truesdale and Mermilliod, 1979) to conduct long-distance migrations to (as juveniles) and from (as reproductive adults) far upstream habitats to coastal waters of the Gulf of Mexico? Other hypotheses have been proposed to explain the occurrence of apparently amphidromous shrimps in the upper Mississippi and Ohio Rivers. It has been hypothesized that larvae from northern populations may have adapted to a more direct form of larval development, which no longer requires saline waters (Gunter, 1978 as cited in Anderson, 1983). On the other hand, Bauer and Delahoussaye (2008) hypothesized that the drainage of brine springs and surface salt deposits distributed around the upper MRS and Ohio River may have made possible seasonal saline "larval nurseries" which supported upstream larval development, negating the necessity of a female downstream migration. Such larval nurseries may not still exist today because of river control and other human impacts on the MR. Yet another hypothesis to be tested is that populations in the upper Mississippi and lower Ohio Rivers may be (or were) ecological sinks, i.e., populations that recruit juveniles produced by populations in the lower MR, but themselves do not contribute to future generations. These juveniles, a result of larval development in the Gulf of Mexico, continued up the MR beyond the location of the parental populations. As adults, they mature, breed and attempt the downstream migration to the Gulf, but the distances from the upper MR to saline coastal waters are beyond the migratory capabilities of adult females and/or their hatched larvae. Such ecological sinks have been proposed for other amphidromous species (McDowall, 2007; McRae, 2007).
These hypotheses should be the focus of future research on the life history of Macrobrachium ohione. Since the 1940s, populations of M. ohione have declined drastically in the upper MRS so much so that only a very small number of shrimps have been collected there in recent times (Taylor, 1992; Barko and Hrabik, 2004). Such small collections of M. ohione illustrate how rare the species has become in the upper MRS, and many researchers now believe that M. ohione may be locally extinct in these areas (Anderson, 1983; Barko and Hrabik, 2004; Bauer and Delahoussaye, 2008). These alarming observations indicate the necessity for a greater knowledge of M. ohiongs life history within the MRS. By obtaining more information on the species where populations are still abundant (lower MRS), conservation and restoration techniques (Bowles et al., 2000; Bauer and Delahoussaye, 2008) can potentially be implemented in the future.
Acknowledgments.--We especially like to thank Cassidy Lejeune and the staff of the Louisiana Department of Wildlife and Fisheries for sampling in both the AR and MR deltas; William Spell at Entergy Gulf States River Bend Nuclear Power Plant for allowing access to its MR property for sampling; and James Delahoussaye for access and help sampling at Butte La Rose. We would also like to thank the undergraduate student workers supported by Louisiana Sea Grant, Courtney Neveu and Joni Yzaguirre, for help in collecting and processing the thousands of shrimps taken in sampling. This study was supported by the Louisiana Sea Grant College Program with funds from the National Oceanic and Atmospheric Administration Office of Sea Grant, Department of Commerce, under the grant No. NA06OAR4170022, Project No. R/SA-04 to RTB and Louisiana State University. Statements, findings, conclusions and recommendations are those of the authors and do not necessarily reflect the views of Louisiana Sea Grant or the U.S. Department of Commerce. We would also like to thank the Louisiana Board of Regents and Southern Regional Education Board for financial support (minority doctoral fellowship) to TJO. TJO also acknowledges financial support from the University of Louisiana at Lafayette (ULL) Graduate Student Organization. This is contribution number 142 of the ULL Laboratory for Crustacean Research.
ANDERSON, G. 1983. Observations on the distribution and movements of Macrobrachium ohione (Smith, 1874) in the Pascagoula River Estuary, Jackson County, Mississippi, U.S.A. (Decapoda, Palaemonidae). Crustaceana, 44:123-140.
BARKO, V. A. AND R. A. HRABIK. 2004. Abundance of Ohio shrimp (Macrobrachium ohione) and glass shrimp (Palaemonetes kadiakensis) in the unimpounded Upper Mississippi River. Am. Midl. Nat., 151:265-273.
BAUER, R. T. 2004. Remarkable shrimps: adaptations and natural history of the carideans. University of Oklahoma Press, Norman.
-- AND J. DELAHOUSSAYE. 2008. Life history migrations of the amphidromous river shrimp Macrobrachium ohione from a continental large river system. J. Crustacean Biol., 28(4):622-632.
BLANCO, J. F. AND F. N. SCATENA. 2005. Floods, habitat hydraulics and upstream migrations of Neritina virginea (Gastropoda: Neritidae) in the northeastern Puerto Rico. Caribb. J. Sci., 41(1):55-74.
BOWLES, D. E., K. AZIZ AND C. L. KNIGHT. 2000. Macrobrachium (Decapoda: Caridea: Palamonidae) in the contiguous United States: A review of the species and an assessment of threats to their survival. J. crustacean Biol., 20:158-171.
CONAWAY, L. K. AND R. A. HRABIK. 1997. The Ohio shrimp, Macrobrachium ohione, in the upper Mississippi River. Trans. Missouri Acad. Sci., 31:44-46.
DINGLE, H. 1996. Migration: The biology of life on the move. Oxford University Press, New York. Oxford.
DUGAN, C. C. 1971. Florida research: freshwater shrimp. Am. Fish Far., 2:8-9.
--, R. W. HAGOOD AND T. A. FRANKS. 1975. Development of spawning and mass larval rearing techniques for brackish-freshwater shrimps of the genus Macrobrachium (Decapoda Palaemonidae). Fla. Mar. Res. Pub., 12:5-24.
GOTELLI, N. J. AND A. M. ELLISON. 2004. A primer of ecological statistics. Sinauer Associates, Inc., Sunderland, MA. 1-510.
GUNTER, G. 1937. Observations on the river shrimp Macrobrachium ohionis (Smith). Am. Midl. Nat., 18:1038-1042.
HARTMANN, G. 1958. Apuntes sobre la biologia de camaron de rio, Cryphiops caementarius (Molina) Palaemonidae, Decapoda. Pesca y Casa, 8:15-28.
HEDGPETH, J. W. 1949. The North American species of Macrobrachium (river shrimp). Texas J. Sci., 1:28-38.
HOLMQUIST, J. G., J. M. SCHMIDT-GENGENBACH AND B. BUCHANAN-YOSHIOKA. 1998. High dams and marine-freshwater linkages: effects on native and introduced fauna in the Caribbean. Conserv. Biol., 12:621-630.
HORNE, F. AND S. BEISSER. 1977. Distribution of river shrimp in the Guadalupe and San Marcos Rivers of Central Texas, U.S.A. (Decapoda, Caridea). Crustaceana, 33 (1):56-60.
HUBSCHMAN, J. H. AND C. BROAD. 1974. The larval development of Palamonetes intermedius (Decapoda, Palaemonidae) reared in the laboratory. Crustaceana, 26:89-103.
IBRAHIM, K. H. 1962. Observations on the fishery and biology of the freshwater prawn Macrobrachium malcolmsonii Milne Edwards in river Godavari. Indian J. Fish., 9:433-467.
JALIHAL, D. R., K. N. SANKOLL AND S. SHENOV. 1993. Evolution of larval developmental patterns and the process of freshwaterization in the prawn genus Macrobrachium Bate, 1968 (Decapoda, Palaemonidae). Crustaceana, 65(3):365-375.
KIKKERT, D. A. AND T. A. CROWL 2009. Upstream migration of amphidromous shrimp in the Luquillo Experimental Forest, Puerto Rico: temporal patterns and environmental cues. J.N. Am. Benthol. Soc., 28:233-246.
LING, S. W. 1969. The general biology and development of Macrobrachium rosenbergii (De Man). FAO Fish. Rep., 57:589-606.
MAGALHAES, C. AND I. WALKER. 1988. Larval development and ecological distribution of central Amazonian palaemonid shrimps (Decapoda, Caridea). Crustaceana, 55:279-292.
MCCORMICK, R. N. 1934. Macrobrachium ohionis, the large fresh-water shrimp. Ind. Acad. Sci. Proc., 43:218-225.
MCDOWALL, R. M. 1992. Diadromy: Orgins and definitions of terminology. Copeia, 1992:248-51. * 2007. On amphidromy, a distinct form of diadromy in aquatic organisms. Fish Fish., 8:1-13.
MCRAE, M. G. 2007. The potential for source-sink population dynamics in Hawaii's amphidromous fishes. Proc. Biol. Hawaii. Streams Estum. Symp., 3:87-98.
POLY, W. J. AND J. E. WETZEL. 2002. The Ohio shrimp, Macrobrachium ohione (Palaemonidae), in the lower Ohio river of Illinois. Trans. Ill. St. Acad. Sci., 95:65-66.
REIMER, D., K. STRAWN AND A. DIXON. 1974. Notes on the river shrimp, Macrobrachium ohione (Smith) 1874, in the Galveston Bay System of Texas. T. Am. Fish. Soc., 103:120-126.
ROME, N. E., S. L. CONNER AND R. T. BAUER. 2009. Delivery of hatching larvae to estuaries by an amphidromous river shrimp: test of hypotheses based on larval moulting and distribution. Freshw. Biol., 54:1924-1932.
TABOR, C. A. 1992. The rediscovery of the Ohio Shrimp, Macrobrachium ohione, in Illinois. Trans. Ill. St. Acad. Sci., 85(3 and 4):227-228.
TRUESDALE, F. M. AND W.J. MERMILLIOD. 1979. The river shrimp Macrobrachium ohione (Smith) (Decapoda, Palaemonidae): its abundance, reproduction, and growth in the Atchafalaya basin of Louisiana, U.S.A. Crustaceana, 32:216-220.
SUBMITTED 9 AUGUST 2010
ACCEPTED 11 MARCH 2011
TYLER J. OLIVIER (1) AND RAYMOND T. BAUER (2)
Department of Biology, University of Louisiana, Lafayette 70504
(1) Corresponding author present address: Department of Biology, University of Louisiana, Lafayette 70504; Telephone: (337) 482-5231: FAX: (337) 482-5834; e-mail: firstname.lastname@example.org
(2) e-mail: email@example.com
|Printer friendly Cite/link Email Feedback|
|Author:||Olivier, Tyler J.; Bauer, Raymond T.|
|Publication:||The American Midland Naturalist|
|Date:||Oct 1, 2011|
|Previous Article:||Indigenous unionid clam refugia from zebra mussels in Michigan inland lakes.|
|Next Article:||Seasonal effects on female fecundity and male sperm availability in a thermally stable temperate population of sailfin mollies (Poecilia latipinna).|