Size, Development of Pigment, Upstream Migration, and Relative Abundance of Young American Eels, Anguilla rostrata in a Coastal Rhode Island Stream

Progressive epide rmal p igmentation of Anguilla rostrata elver s in a coastal Rhode Island stream was essentially identical to that de scr ibe d for ~ anguilla, and proceeded rapidly in fr eshwater. Earlier arriving elvers averaged larger than those arriving later in the season, paralleling studies of European elvers. Mea n tot al lengths of elver s collected in 1984 wer e significantly greater than those for 1983 and ar e the lar ge s t re corded for t he Western North Atlantic. Thes e differ ences may be due to sampling technique. Migration of e lv ers into freshwater may be i nduc ed by decreasing flow rates and /or increasing stream temperature s . The relatively slow upstream migratio n of elvers in the lower sect io n of the stream (a pproximately 200 m per month) is attributed to high stre am gradient and t he presence of obstructions. By late summer and early fall, elvers (age I+) had acquired t he coloration typical of yellow eels and had grow n 20 t o 30 mm larger than elvers arriving in freshwater in late winter and spring . The frequency of II+ and older eel s increased with incr e asing d i s ta nce from the tidal zone. Th is indi ca tes that the upstream mi grati on of e lvers is limited, and that the colonization o f inland wate r s is accomplishe d mainly by yellow eels in their second and later years of continental life.


Introduction
In the late winte r an d spring, elvers of the American eel, ~ rostrata, move from estuaries towards mouths of freshwater stre am s. During this phase of their migrat i on elvers gradually acquire epidermal pigmentation and asc e nd flowing waters, negotiating obstructions to their progress by various means. Elvers then begin to acquire the coloration characteristic of the yellow eel, which is the primary growth phase. However, little is known about upstream movements of ~ rostrata beyond a few anecdotal observations . No previous investigation has specifically addressed upstream movements of elvers of this species and related changes in their morphology, although considerable work in this area has been performed for the European eel and some South Pacific species.
The works of Grassi (1913) and Schmidt (1906) provided the fundamental studies of development and distribution of leptocephali and glass eels of the closely related European eel, Anguilla anguilla, which formed the basis for most later studies of early Anguilla life histories. Schmidt's work, and also that of  established a detailed pigmentation classification for developmental stages of early leptocephalus to early yellow eel. Strubberg (1913) used a somewhat simplified version of Schmidt's and Gilson's classifications in shipboard experiments to investigate changes in elver pigmentation associated with various environmental factors. Strubberg found that the rate of pigmentation in elvers increased with water temperature, but was not significantly affected by salinity or incident light.  used a pigmentation classification modified from Strubberg's for both~ anguilla and ~ rostrata elvers present in his collections from Denmark.
Boetius did not note differences in pigmentation between the species.  extended Strubberg's experimental conclusions concerning pigmentation and water temperature, noting a pigmentation rate increase in ~ anguilla elvers entering warmer fresh water as opposed to delayed pigmentation in elvers from colder coastal water. Strubberg (1913) also noted a reduction of lengths of ~ anguilla elvers with the development of pigment, and cited time and/or temperature as the probable causative agents. Similar conclusions have also been drawn by others (Grassi and Calandruccio, 1897;Schmidt, 1906;Menzies, 1936; However,  and  attributed the observed decreases in length to later arriving elvers being shorter. Boetius theorized that these elvers metamorphosed in an area more distant from freshwater than those that had arrived earlier, and thus had a longer migration. Vladykov (1966)  Because the timing of elver run peaks may also vary from year to year , env ironmenta l influences may play a ma jo r role in t he distribution of leptoc ephal i and elvers. Smith (1968) described a collection of elvers take n on 19 January in Flor ida , and suggested that the ascent of fr eshwater in Florida occurs earlier than it does farth er north.
Invasion of freshwater by ~ rostrata is thought to be similar to that of A.i. anguilla, for which many studies of upstream migration exist. Upstream movement of elvers by tidal transport in estuarine areas has been shown for European Creutzberg, 1958;McCleave, 1980) and also American (McCleave and Kleckner, 1982) eel elvers. Deelder (1958) provided descriptions of changes in migratory behavior and response ·to salinity in A.i. anguilla elvers once they had arrived at a boundary area between the outlet of a stream and a tidal area. Factors affecting the movements of elvers into streams can be complex. Sorensen (1951) stated that upstream migration of A.i. anguilla elvers was initiated when a stream reached a certain threshold temperature of approximately 15 C. Other cues which appear to initiate upstream migration in elvers of various species of Anguilla have included lunar cycles (Gollub, 1959;Jellyman, 1979), rainfall and/or stream size (Jellyman and Ryan, 1983), intensity of daylight and current velocity (Sorensen, 1951;Sloane, 1984aSloane, , 1984b, and high spring tides (Matsui, 1952;Sorensen, 1984).
Progress of elvers upstream can be rapid. A.i. anguilla elvers have been observed to move upstream at the rate of 1 to 3 kilometers per month . However, their upstream migration can also be inhibited by environmental conditions. McCleave (1980) determined that A.i. anguil.lii elvers could not progress against currents greater than 50 cm per second, and inferred that movement upstream was facilitated by elvers alternating between burst swimming and utilization of bottom topography (i. e. resting behind rocks, boulders, etc.) or burrowing into bottom sediments. Dams and other man-made structures, as well as natural ones, also impede the upstream progress of elvers and young eels .

I. Pigmentation Qf_elvers
The epidermal pigmentation of Annaquatucket elvers was very similar to that desc ribed for ~ anguilla. The ear l iest -arr i ving elvers all possessed the head pigmentation or "tache cerebrale" of European elvers entering freshwater.
Caudal pigment wa s always present, and no elver was wholly unpigmented. A pigmentation regime of seven stages was established so that the entire spectrum of the pigmentation range of the col lected sample would be represented, a nd eac h stage could be recognized with a minimum of discrepancy between stages (Fig. 3). Comparisons of this regime with other historical classifications are given in Table I Epitidal stations to the Fishway station is presumed to have occurred between these two dates. This same time pe riod coincides with the gre atest increase in stream temperature (from 10.5 to 21.5°c) and the lowest water levels (5 or more cm below the datum) for th e sampling period. Relative abundances of elvers a t t he Tida l and Epitidal stations decreased to less than 2.5 pe rcent after 4 June, wh ile the relative abundance of elvers at the Fishway station declined to 13 percent after this date. Stage 7      in 1984 occurred between 23 April and 7 May (Fig. 5), when stream tem perature increased at its highest rate for the season (Fig. 8). Clearly, there is a strong association between stream temperature and the rate at which elvers recruit pigment. In addition, later arriving elvers tend to be more pigmented, which suggests that initiation of pigmentation takes place further offshore as the season progresses and the temper a ture of coastal waters increases. Reduction in elver mea n total length over time also reiterates previous studies with A..,__ anguilla. Since development of pigment is not highly correlated with length, it is likely that this trend of decreasing length is due to smaller sizes of later-a rriving elvers rather tha n a metamorphosis or "shrinking" of individual fish.
Differences in mean total length of elvers between collecting statio ns are not evident (Fig. 6), sinc e there is great variability in sizes of elvers within any single collection, and the reduction in size is fairly gradual as the season progresses. Also, it is not likely that individual elvers remained in a particular sampling area for any great period of time; thus the collections represent successive groups of elvers migrating upstream.

Mean total lengths of elvers from pooled collections
for each year are unex pectedly dissimilar to published data.
Although the 1983 collections consist of elvers of the size that one would expect from Vladykov's (1966) report of sizes of elvers taken fr om the Northwestern Atlantic (Fig. 7), elvers from the 1984 collection averaged longer than Vladykov's specime ns from Nova Scotia, which were the largest from his series. So me of Vladykov's collections were stored for up to eight years before they were measure d ; therefore his estimates of mean total length may have been affected by shrinkage. The much greater mean total length of elvers caught in 1984 versus 1983 may be due to sampling selectivity, since elvers were captured by dipnetting in 1983 and by electrofishing in 1984 . Samples taken in 1982 by J. DiCanzio were also collected with a dipnet, and are comparable in size to 1983 elvers (Fig. 7).
The two collections of this study (1983 and 1984) may represent a true natural difference in size of elver s from the two different year classes, but there is strong evidence for sampling bia s betw e en the two methodologies.
Arrival of elvers and their subsequent movement upstream also appear to be regulated by environmental factor s . Su ch facto rs ar e , however, not entirely limiting. This is evidenced by the collection of a 57.8 mm elver in pigmentation stage 5 below Mi l l Pond on 6 November, 1983.
Such ~n appearance of elv e rs "out of season" indicates that they may a rrive on the coasts of North America throughout the year, yet only in su bst a ntial numbers during the spring months. Sloane (1984a) observed similar year-round arrivals of~ australis elvers in Ta sm anian streams.
Although the r a t e of upstream migration could not be precisely "defined, the main concentration of elvers seems to take about one mont h to move from the Tidal and Epitidal stat i ons t o th e Fi s hway station. In comparison, Sorensen (1984) postulated that it took elvers several weeks to pass a 400 m s tretch of Gilbert Stuart Brook, which has a gradient similar to that of the Annaquatucket. Upstream movements of elvers may be mediated by temperature and/or wate r level. Tempe r at u r e s greater than 10°C have been noted to initiate upstream migration of elvers (Sorensen, 1951;Smith, 1955;Ma nn, 1963). This generalized "threshold" temperature correlates closely with the temperature increase associated with the greatest amount of movement of elver s from the Tidal to Epitidal statio ns (from 10.5 to 21.0°C). Decreased water levels (and thus decreased current velocities) may encourage upstream movement of elvers which had previously been hindered in their progress at the Tidal station by heavy spring runoff. However, some elvers were found in areas that had been flooded above the stream banks by high water levels. Under these circumstances, elvers may facilitate their upstream progress by moving through these flooded areas away from the main current. Thus water level may affec~ elver movement upstream positively or negatively depending on habitat type (low banks · which become flooded versus high banks which do not). Elvers migrating from the tidal zone may also time their ascent based on spring tides or lunar cycles (Jellyman, 1979;Sorensen, 1984), but these effects are difficult to correlate precisely with the results from this study since samples were taken only once every two weeks.
The morphological changes in A_,_ rostrata elvers associated with upstream migration, as well as the effects of stream temperature, gradient, and water level are comparable to those described for other species of Anguilla.
The similarity between species at this particular life history stage is not surprising, since the environmental conditions that elvers encounter in migrating upstream are about the same for most streams and rivers that these species enter as juveniles. However, the rate at which ~ rostrata elvers ascend the lower section of the Annaquatucket is much less that that reported in most cases for ~ anguilla. This phenomenon may be due to the high gradient and current velocity of the Annaquatucket in this lower region. These conditions are in direct contrast to those found in larger rivers with lower gradient and current velocities (especially near banks) where extensive movements of elvers have been observed. Tidal transport of elvers in larger rivers may also account for a large portion of the distance that elvers migrate upstream in these systems.
Thus, the physical nature of a stream or river and the specific environmental conditions that elvers encounter while migrating upstream may dictate the speed and distance of their ascent.   LaBar and Facey, 1983) and arrival of eel s of age III+ or older at locations far upstream (e.g. Lake Champlai n and Lake Ontario ) provide the only information regarding monitored movements of yellow stage A.... rostrata i n freshwater.
Knowledge of upstream migration in juvenile European eels is also far from comp lete. While European elvers have been known to migrate from 30 to 150 kilometers during their first year in fre shwater (Tesch, 1965;Deelder, 1970;, movements of juve ni le eels past the elver stage have never been intensively studied. Tesch (1967) collected primarily juvenile fl_._ anguilla in weirs more than 150 Km upstream ·on two German rive rs . Elvers at these weirs we re relatively rare . Tesch assumed that because these juvenile eels averaged 150 mm in total length, they had taken at least two years since entering freshwater to reach these upstream sites.
Other s tudie s (Deelder, 1970;Moriarty , 1978) have shown similar limited mov eme nt o f fl_._ anguill.a yellow eels during their first few yea rs in freshwater, with eels collected farther inland showing an increase in size and age. Such effects on the distribution and age composition are thought to be amplified by obstructio ns to eel progress, such as weirs and dams (Tesch, 1977). Sloane (1984) postulated that A_,_ fil.!Stralis past the elver stage migrate upstream for several years in succession , since age groups II+, III+, and IV+ were more abundant than elvers at sites furt her up s tream. No elvers of thi s species were collected more t han 6 Km inla nd.
This study investigates the distribution of A~ rostrata elvers and yellow eels smaller than 250 mm TL at three sites in a coastal Rhode Island stream. Relative proportions of age classes at each site are used to determine the extent of upstream migration of elvers and young yellow eels.

Materials and Methods
All specimens used i n this study were taken from the Annaquatucket River watershed, Washington County, Rhode Island (41° 33' N; 76° 26 1 W; Fig. 1 After the electrofisher had reached the net, the net was removed from the stream and fish were extracted by hand.
Total lengths were determined to the nearest millimeter using a measuring board, and eels greater than 250 mm total length were returned to the stream. All three sites were sampled on 20 August, 23 September, and 6 November, 1983.
The Featherbed and Belleville sit es were sampled again on 7 August, 1984. After transport to the laboratory, eels were anesthetized with MS-222, me asured to the nearest mm, labeled, and frozen for later analysis.
Otoliths (  avoid crushing sma ll otoliths, a grinding apparatus similar to that described by Neilson and Geen (1981) was used on otoliths from eels smaller than approximately 100 mm.
Ground a nd polished otoliths were covered with a drop of mineral oil to prevent dry i ng while they were s tored until they could be read.
Otolith annu li were identified and counted using a dissecting microscop e at 100X. Trans mi tted light was usually sufficient to re veal annuli, but reflected light wa s also used on each otolith to reveal transparent and opaque r ings d escr ib e d by Gray and Andrews (1971) that are associated with annuli. The sea-water or metamorphic ring was regarded as one year of larval growth (I+) and all subsequent rings as later annuli (II+ and older). As a chec k for inaccuracies in aging, a linear discriminant analysis (LDA) wa s performed on the aged fish using total length as a me an s to cl assify into a ge groups . If the observed age wa s no t equal to the age determined by th e LDA, the otolith was read again t o check for an aging e rror . The reage d data was th en subjected to a second LDA to improve the agi ng classi fi c atio n .  (Fig. 28) showed a greatly expanded outer grow th zone compared with those taken in the spring ( Fig. 2A). Summer and fall elver otoliths were also characterized by a jagged margin and prominent rostrum, while spring elver otoliths had a more uniform periphery.
Age II+ eels were distinguished from elvers by the presence of an annulus surrounding the metamorphic or seawater otolith ring. Annuli of all eels were normally accompanied by opaque rings (Fig. 28

Discussion
The d istribution of age groups at the three sampling sites (Fig. 3)  eels with limited ho me ranges within a stream or lake (Gunning and Shoop, 1962;Helfman et al., 1984;Labar and Facey, 1983), most eels probably stop migrating upstream at some point in their lifetimes, but it is not clear at what age they do so. The c hanges in morphology of elvers during their upstream migration can be particularly dramatic.
Elvers arriving in late winter and spring are initially semi-transparent, and accumulate pigment rapidly once they enter fr e shwater. However, elvers which attain "full" pigmentation (stage 7) by late spring are quite different in appearance from elvers collected in the late summer and fall; the latter being larger and more yellowish or greenish. Elver growth during the summer is also evident in the expanded transpar e nt g rowth zone of the otolith margin (Fi g . 2B). Absence of annuli and/or opaque zones near the margins of otoliths from late summer and fall eels suggests that annulus formation occurs during the winter, which is in agree ment with .
The conclusions drawn so far in this study contrast many reports re garding the European eel, which describe l e ng t h y migratio n s of elvers upstream for several kilometers or more. Clearly, e lvers in the Annaquatucket progress up s tream slowly, and their numbers become much reduced at upstream sites in co mparison to their relative abundance in the tidal zone. This disparity between An naquatucket elvers and those from other systems which have little difficulty moving large distances upstream can probably be related to a fundamental difference in habitat types. Many European studies describing extensive elver migrations have taken place on large river systems (i.e. Deelder, , 1958Deelder, , 1970, where current velocities are probably not great during the spring months in comparison with the Annaquatucket. Most of these studies do not provide information regarding what proportion of these migrations (in terms of dista nce) are facilitated by tidal transport, whic h could be a major factor of elver movement in larger rivers. Progress upstream by elvers in such systems is also likely to be e nhanced by the lack of current near river banks; areas in which elvers have been noted to perfo rm most of t he ir migratory activity (Tesch, 1977).
This habitat type i s in direct contrast to the high gradient of the lower Annaquatucket system, which experiences great fluctuatio ns in current velocity in response to changes in local precipitation . Fishways on the Annaquatucket seem to hav e little effect on assisting the upstream migration of elvers. In fact, they may be inhibitory in this respect in that they serve as attractors to elvers seeking flowing fresh water, yet elvers may have difficulty ascending fishways when current velocities through the baffles are high. Other studies in Europe have shown a similar inability of e lvers to ascend fishways with high current velocities (T esch , 1977). The physical nature of elvers and their specific behavior during migration may exclude the m from utilizing the fishways in a way that other species of upstream-migrating fishes do. Tesch (1977) has suggested that only l arger ee ls are capable of ascending sections of streams which posess many dams, weirs, and other obstructions . It is possible that migrating elvers wait until water levels and current velocities drop before ascending fishways, or find other means of circumventing da ms, such as climbing the wet concrete structures of the dams themselves. Also, the effect of standing bodies of water (i.e. lakes and ponds ) on the progress of elvers throu gh a river system is unkn own. Elvers have been reported to wander rando mly in still waters (Tesch, 1977), and thu s would be expected to have difficulty locating an inlet stream once having entered an impoundment.
In spite of all the potential restraints on their upstream mig ration, elvers are able to pass above the Mill Growth o f older eels may no t be accurately described f rom these estimates.