Abstract. The spatial distribution of Eurytemora americana in the Duwamish River estuary was investigated in April, 1996. The copepod occurred chiefly in the marine layer (S>25 PSU) and was most abundant 7.5 km upriver from the mouth of the west waterway (7.5 Rkm). The population diminished upstream to the toe of the salt wedge (10 Rkm) where the extent of the marine layer ended; downstream, the population diminished at 3.6 Rkm, where competition with marine copepods may control the zooplankton assemblage (3.6 Rkm). Although the existence of a permanent marine layer in the Duwamish River allows E. americana to persist, other mechanisms such as vertical migration on a tidal cycle scale or migrational movements toward physical cues may also be used for retention.
 
 

Introduction
 

The Duwamish River estuary is a temperate, tidally influenced system supporting a wide variety of life (from salmonids and Dungeness crabs and other shellfish to the higher trophic levels of birds and marine mammals). The river is lined with industrial and commercial facilities and has been channeled and dredged to support a large shipping industry. Many pollutants have been introduced into the estuary via sewage discharge, local industry, and non-point sources. In the past, it has been one of the most polluted rivers in Washington State and has required massive undertakings to assess the extent, sources, and implications of the toxicants in the river [METRO, 1983]. For these reasons, it is imperative to understand the fate of toxicants within the estuary and, subsequently, the distributions of possible toxicant vectors such as zooplankton.

The river has a dynamic, two layer salt wedge that extends from the mouth of the river to as far as 19 km upstream from the mouth of the west waterway (19 Rkm) [Dawson and Tilly, 1972]. The toe of the wedge may only have a daily movement of less than 5 km, the annual mean being at ~ 12 Rkm [Warner and Fritz, 1995]. The excursion distance of the salt wedge depends primarily on the tidal height of Puget Sound, into which the Duwamish River flows, and the freshwater flow rate from upriver. The salinity extremes extend from marine values of S>28 PSU to S=0 PSU. The null point and high-salinity gradient zones are regions of elevated chemical reactions and flocculations of sediments and metals [Morris et al., 1978], and therefore a regime in which toxicants would be of prime concern.

Even though the habitat is very dynamic in this estuarine environment, the planktonic calanoid copepod Eurytemora americana occurs there, [Frost and Cordell, personal communication, 1996]. Estuarine copepods have been shown to reside in various salinity regimes [Miller, 1983] and seem to be hardy animals that can persist local extinction in estuarine areas of pollution and extreme change [Soltanpour-Gargari and Wellershaus, 1987]. For these reasons, if E. americana resides within the turbidity maximum as suspected, it would be a good choice for the study of toxicants that may be prevalent throughout the high salinity gradients and consequent turbidity zone.

E. americana was first described by Williams [1906] and later taxonomic literature provided by Heron [1964]. Although extensive study has been done on Eurytemora affinis and to a lesser extent on Eurytemora herdmani, only qualitative results [Bousfeild et al., 1975] have been published on the physiology, ecology, or distributions of E. americana except for those by Miller [1983].

An item of interest is how E. americana is retained within the estuary. Many estuarine copepods are suited to wide salinity variations, rapid reproduction (upwards of 40 eggs d^-1), and the production of resting eggs to be produced in unfavorable conditions. Whether or not E. americana migrates to stay within its primary habitat is unknown, though Morgan [1993] and Hough and Naylor [1992] have produced evidence of Eurytemora affinis vertically migrating downward on a tidal cycle (12.4 hours) to decrease losses due to advection. Although none of these studies has been done on populations of on E. americana to date, B. Frost [personal communication, 1996] has suggested that the animal may use migration to possibly increase its retention by traveling from the outflowing freshwater current to the upriver flowing saltwater component of the system. Other studies have shown that the entrainment alone of E. affinis in the resultant salt wedge currents can account for the retention; the copepods act as particles in the system without vertical migrations [Soltanpour-Gargari and Wellershaus, 1987; Castel and Veiga, 1990].

The possibility of E. americana's residence below the high-salinity gradient zone must also be considered, for the copepod may tend to stay within the upriver flow of the lower marine layers, and occasionally get mixed up and out of the to maintain position in the estuary. Miller [1983] and A. Liljevik [personal communication, 1996] have observed E. americana clinging to the sides and bottom of containers in the laboratory. Vertical migration may also play a role. E. americana may move downward whenever cues such as lower salinity are detected. In this situation, the highest densities of the animal would be found in the lower marine layers, although this leads to the question of where this copepod's limitation in the lower end of the river exists and why, for the same salinity extends into Puget Sound where E. americana does not exist [B. Frost, personal communication,1996].

A final item of interest is the type of estuarine environment in which E. americana or other resident estuarine copepods can exist. We know from presampling that the Duwamish River estuary can support this biota, but why can other estuaries in the Puget Sound region, such as the Snohomish River estuary, not support such resident zooplankton [B. Frost, personal communication, 1996]?

In this study, the distribution of E. americana is examined within the Duwamish River estuary between the upriver limit of the salt wedge and fully marine waters. It is suspected to reside in the turbidity maximum zone between the fresh and marine water layers throughout the salt wedge (in the high salinity gradient layer) or in the lower marine layer. I attempt to compare the copepod distribution to the salinity structure, the turbidity structure, current profiles, and depth in the estuary.
 

Methods
 

Since the salt wedge is a dynamic feature of the river and the copepod concentration maximum is presumed to move in relation to the salt wedge, methods of sampling were executed while the wedge was moving and while the wedge was stationary. In order to determine the distribution of E. americana with relation to the salt wedge layers, two field sampling patterns were executed: one sample pattern (survey) was done around a slack high tide when the salt wedge was presumed stationary; the Eulerian pattern was conducted as the wedge developed at a single location.

The Survey mode consisted of nine sample sites along the main river channel, with samples taken at approximately every 1 km from 50 meters upstream of the salt wedge toe (11.5 Rkm) to 3 Rkm (Figure 1). The sampling was performed around slack high tide plus or minus 135 minutes on Tuesday, April 2, 1996 for stations 1 - 6; stations 6 - 10 were sampled on Wednesday, April 3, 1996 with station 6 sampled twice over the two days. At each sample site three samples were taken: one at the surface (S<5 PSU), one in the layer of maximum salinity gradient (10 PSU>S>25 PSU), and one within the marine layer (S>25 PSU). At stations 1 and 2, where the depth limited the number of samples, only two samples were taken at S<0 PSU and S>18.5 PSU. The depth of each sample was dependent upon the vertical location of the salinity layers. Samples were taken only in the main channel.

A preliminary survey at low tide was also completed on Monday, April 1, 1996 to find the salt wedge toe at low tide and determine the excursion distances of the salt wedge toe at present condtions. Salinity and temperature were recorded every 0.5 m with a YSI temperature-salinity (T-S) meter.

The Eulerian sampling study was completed on Wednesday, April 3, 1996 while the salt wedge developed on a flooding tide at a single location (Station E in Figure 1). Sampling was done approximately every 30 min from 12:25 hours to 14:50 hours (approximately 2 h after low tide to 2 h before high tide). The intent was to sample from the period before the encroachment of the salt wedge toe to a time when the salt wedge profile was fully developed (i.e. the marine layer is at least 2 m thick). Each sampling consisted of a conductivity-temperature-depth (CTD) cast and three zooplankton sample pumpings at depths of 10 cm, 2 m, and 4 m.

Figure 1. Sampling stations in the Duwamish River estuary. Redrawn from Coastline Extractor, USGS, 1996, via internet.

Sampling was done from a small motorboat (22' WeeLander Beachmaster) outfitted with DGPS for navigation. Zooplankton samples were taken with a submersible pump rated at 2000 gal hr^-1 and packaged with an Aanderra current meter and YSI T-S meter electrode to assure pumping in the appropriate salinity layer. In addition to using the YSI T-S meter, a recording Seacat Sea-Bird SBE-19 CTD coupled with a 5 cm transmissometer was cast to profile the water column for salinity, temperature, depth, and turbidity. The pumping rate was 20 L/min and was not significantly affected by depths encountered (<12m). A volume of 80 liters was pumped from depth for 4 min and filtered through a 100 m filter. The catch was fixed with 10 percent formalin/seawater mixture in a 250 ml jar. In addition to pump sampling, net tows were done at various locations and depths corresponding to various pump samples to serve as a check for sampling error such as evasion. The net used was 0.5 m in diameter, with a mesh pore size of 335m and was equipped with a TSK (Tsurumi-Sikie-Kosakusho Co.) flowmeter.

In the laboratory, the samples were analyzed by counting E. americana stage C6 female copepods to determine population density. An "other copepod calanoids" group was also analyzed and consisted of Acartia spp. and Psuedocalanus spp. This "other" group was analyzed in the same manner. Nauplii and younger copepodite stages were not counted, although very high or very low densities were noted. Lab counts were performed with the use of a dissecting microscope. Pumped samples were counted in whole; net tow samples were subsampled twice at 3 percent of total count
 
 
 
 

Results
 

Survey Copepod Populations

According to the data shown in Figure 2, E. americana was found to have a maximum population density around station 6, tapering off in both upstream and downstream directions. The copepod was found to reside mainly in the marine layer (S>25 PSU), occurring sparsely in the intermediate salinity regimes (10 PSU<S<20 PSU). E. americana was not found in the fresh water layer. Other calanoid copepods extended from maximum densities at the mouth of the river to low densities at station 1 (Figure 3). These copepods were also most prevalent in the marine layer and present to a lesser extent in the intermediate salinity layer.

Figure 2. Densities of E. americana C6 females at three different salinity regimes in the Duwamish River.  

Although evasion was expected to be encountered, the pump sampled the zooplankton adequately, for there was no significant difference in population densities between the net-towed and the pumped samples taken at the same times and depths.
 

Physical Properties on the Survey

The river was a strongly stratified two layer system with a fresh water top layer no thicker than 1.5 m at high tide (Figures 4a-4i). The intermediate layer was fairly thin with the salinity gradient extending from S=5 PSU to S=25 PSU over no more than 2m and extended upriver to past station 1. The thickness of the marine layer depended upon the tidal stage and the depth of the river, which ranged from 12.5m at station 9 to 2m at station 2 during high tide (Figure 5a). The marine layer was not fully developed at station 1 (11.3 Rkm), although salinities of 24 PSU were obtained near the bottom.

The preliminary survey at low tide on Monday, April 1, 1996 revealed that at the low tide, the water column at station 2 was completely fresh water (S<2 PSU). At station 3, the salt wedge was fully developed with fresh water in surface layers and salinity values of S>25 PSU near the bottom (Figure 5b).

There was a turbidity maximum in the water column throughout the river at 1-3 m depth. This coincided with the boundary between the lower limit of the salinity gradient layer and the top of the marine layer. The depth of this interface is deeper upstream due to the thicker fresh water layer. The temperature profile shows a higher temperature surface layer (9.5-9.9 C) than the marine layer (8.5-8.6 C) due to irradiation. The CTD profiling was done in the afternoon, after the maximal amount of surface heating could occur. The vertical temperature gradient is negatively correlated with the salinity gradient.

Although rough current measurements of the different layers were conducted, no correlations were seen, for at the high tide all layers were flowing downstream with the occasional eddy denoted by upstream signals. The current speed was variable and could not be distinguished between all layers at every location.
 

Figure 4 (not shown). CTD profiles of (a.) station 1, (b.) station 2, and (c.) station 3 during survey.

Figure 4 (cont) (not shown). CTD profiles of (d.) station 4, (e.) station 5, and (f.) station 6 during survey.

Figure 4 (cont) (not shown). CTD profiles of (g.) station 7, (h.) station 8, and (i.) station 9 during survey.

Figure 5 (not shown). The salinity contour map of the Duwamish River at (a.) low tide and (b.) high tide from stations 1-9 (5-12 Rkm). Contour lines are plotted in increments of S=2.5 PSU. This figure shows that the marine layer is persistant throughout the tidal cycle at station 3.
 

Eulerian Copepod Populations
 

During the Eulerian sampling, the population density of E. americana at the 4m depth was very high near low tide and diminished as mid-tide levels approached. Between mid and high tides, the density remained constantly lower than at low tide. E. americana was not found at the 2m depth through mid-tide, but low densities were observed as high tide approached. E. americana was not found in the surface layers (Figure 6).

Figure 6. Distribution of E. americana C-6 females at station E around estimated maximum flood current (13:43 hours). Depths of 0m, 2m, and 4m were sampled. Low and high tides occurred at 10:35 hours and 16:50 hours respectively.

Densities of calanoids other than E. americana were distributed chiefly in the marine layer in high densities throughout the sampling period, sparsely in the intermediate depths; they were non-existent in the surface layers (Figure 7).  

Figure 7. Distribution of female calanoid copepods other than E. americana at station E around estimated maximum flood current (13:43 hours). Depths of 0m, 2m, and 4m were sampled. Low and high tides occurred at 10:35 hours and 16:50 hours respectively.

Eulerian Physical Data
 

Light transmission varies with respect to time at station E (Figure 8). Around mid tide (at the estimated flood current maximum) the turbidity increased at the 2m depth, while it decreased at the 4m depth. The turbidity at the surface remained constant.

As shown in Figure 9, the salinity values at all depths remained reasonably constant during the sampling period. Temperature was also constant with time for each of the three depths. The turbidity peak rose in the water column from 4m depth at 12:44 hours to 2m depth at 15:04 hours.

Figure 9. Values of salinity at station E before and after estimated maximum flood tide (13:43 hours). Depths of 0m, 2m, and 4m were sampled. Low and high tides occurred at 10:35 hours and 16:50 hours respectively.
 

Discussion

Figure 2 clearly represents an area of E. americana residence between station 2 and station 9 and with the center at around station 6. The population also resided primarily in the marine layer (S>25 PSU), suggesting that the occurrence of E. americana is not related to the turbidity maximums that occur above the marine layer, but to one or more other physical characteristics of the marine layer at high tide. Conversely, at mid tide, this is assumed to occur due to the elevated mixing occurring at the maximum tidal flood at station E in particular. Station E is an area on the edge of a shelf break (see bathemetry of Figure 5a and 5b) and can constrain the deeper marine level inhabitants to flux up into the areas of mixing.. The survey stations do not show this correlation, for the current regime at slack current (high tide) does not allow for this; the marine layer at high tide is above the level of constriction at station E. Although Castel and Veiga [1990] present data that suggest that the turbidity maximum in estuaries correspond to higher densities of Eurytemora affinis, their study used scales of large magnitude (>10 km between sampling sites) which could result in a spatially incorrect model for the Duwamish as far as turbidity is concerned.

Salinity has been suggested as a distribution constraint for a number of copepods [Bousfeild et al., 1975 and Miller, 1983] including E. affinis [Solanpour-Gargari and Wellerhaus 1987]. E. affinis has been found in estuarine salinity regimes from 0 to 35 ⁰/₀₀ [Von Vauple-Klein and Weber, 1975; Roddie et al., 1984] and even in inland fresh waters [Saunders, 1993]. This widespread occurrence and salinity tolerance in E. affinis may provide insight to the survivability of E. americana and its distribution in relation to salinity. A. Liljevik [personal communication, 1996] determined the effect of salinity on the survival of E. americana from the Duwamish River. These data indicate that the copepod cannot survive in fresh waters (0 PSU), but can survive in salinities from 10 PSU to as high as 33 PSU. The fact that E. americana prefers higher salinities corresponds with the marine layer population distribution found in the Duwamish River estuary.

Although salinity may seem to play a major part in the distribution of E. americana in the estuary, depth can not be dismissed. Morgan [1993] found the abundance of E. affinis, an epibenthic copepod, to be concentrated at greater depths, regardless of salinity. From this study, it is not known if depth has any part in E. americana's distribution, for only one sampling depth was taken within the marine layer.

Temperature in the river is determined by the irradiation of the surface waters, the intrusion of the marine waters, and the mixing of the two. The temperature profiles (Figures 4a-4i), show that the surface water temperatures are greater than those of the marine layer. Although this is not true for colder weather conditions, when temperature is homogeneous with depth, the same pattern in E. americana is seen [personal observations]. In addition, the observed change in temperature of the water column did not exceed 2 C. Therefore, with the same distribution among the population and differing temperature profiles, and little temperature difference, temperature can be dismissed as an attraction cue.

Although current direction or speed are other possible factors the copepod may cue upon, no relationship was seen between current speed or direction and population densities of any copepod species. The current speed data were too variable and the current direction data were for the most part constant regardless of population densities.

The retention of E. americana could be a result of a number of different mechanisms. Firstly, since the salt wedge does not have a great excursion between low and high tide (Figures 5a and 5b), the copepods' distribution is not thought to migrate laterally with any great extent (i.e., over 3 km) on the temporal scale of a tidal cycle. Therefore, the population within the Duwamish can resist expulsion from the river while staying within the marine layers, unlike a river which is totally flushed with fresh water at high tide. As B. Frost [personal communication, 1996] found, a representative river in the Puget Sound area is the Snohomish River which had no resident calanoid copepod populations due to its entire flushing (Figure 10). Note the lack of a marine layer (S>25 PSU) present within the Snohomish River at low tide.

Figure 10 (not shown). Salinity contour plots of the Snohomish River at (a) high tide and (b) low tide. The fresh water entirely flushes out the river at low tides, precluding the occurrance of endemic zooplankton.

Another possible means of retention is the vertical migration of copepods on a tidal cycle. In this case, the copepods would primarily reside within the marine layer and migrate upwards to feed, either on phytoplankton or particle attached bacteria [Baross and Crump, 1994] between low and high tides on the flood current. Phytoplankton, at relatively low levels in the river [ J. Zyskowski, personal communication, 1996] is an unlikely major source of nutrition; particle attached bacteria is a more likely nutrition source. The greatest abundance of this particle attached bacteria is presumed to be most abundant in the high turbidity zone [J. Thomas, personal communication, 1996] layered just above the marine layer and correlated to the lower light transmission in Figures 4a - 4i. As high tide approaches, the copepods would have to retreat into the marine layer to be kept from washing out of the river on the ebb current in the mixed and surface layers. Once in favorable conditions, the rapid and proficient reproduction process could compensate for the losses due to advection in the surface layers. Although no studies have been published to date on E. americana with respect to this (or any other) mechanism of retention, E. affinis has been shown to follow such patterns in the Columbia River [Morgan, 1993] and in the Conway estuary of Wales [Hough and Naylor, 1992].

Castel and Veiga [1990] refute the vertical migration theory with respect to E. affinis, instead postulating that the copepod behaves as a passive particle. They base this theory upon the fact that these copepods were related directly to the turbidity maximum in the river, but they did not consider the effect of depth, which would be very important in a two-layer flow estuarine system. Morgan [1993] argues that if Castel and Veiga [1990] considered depth and tidal stage rather than an average population value over the entire water column, they could have postulated otherwise. In addition, the data obtained outside severe mixing events (Eulerian sampling) in the Duwamish River, do not show any direct relation between the distribution of E. americana and turbidity.

The distribution of E. americana therefore appears related chiefly to the salinity structure, the copepod staying within the marine layer, except upon forced upward flux during intense mixing events. Although the upstream limit can be seen to be associated with lower salinity levels between stations 1 and 2, the downstream reduction of E. americana is not accounted for by physical means. Species competition may account for this trend. Figure 11 shows the distributions of both E. americana and other calanoid copepods, which are marine in origin and entrained into the estuary. The upstream distribution of these marine copepods appears limited by lower salinities (Figure 3 and 11). A niche overlap can be seen in Figure 11 around station 6. As the marine copepods start to decline in abundance, the E. americana population tends to peak. This is a classic feature of a system with overlapping niches and competitive species.  

Conclusions
 

The distribution of E. americana was centered around station 6 and extended upriver to the toe of the salt wedge and downriver to past station 9 and is mainly limited to the marine layers (S>25 PSU). The copepod, having a low tolerance for fresh water [A. Liljevik, personal communication, 1996] is limited upstream by lower salinity levels. Downstream, the a negative correlation is seen with the increase in marine calanoid copepods. A limiting factor for the E. americana abundance suspected here is competition. Although a correlation is seen with respect to turbidity at Station E, E. americana is not presumed to be limited by or characterized by a turbidity maximum.

Possible mechanisms for retention of this copepod population may include vertical migrations on a tidal cycle, epibenthic resistance by means of staying within the net upstream movement of the marine layer, rapid reproduction rates to populate faster than advection out of the estuary, or a combination of any of these methods.
 
 

Further Studies
 

Questions that may be answered by further study of the Duwamish River estuary E. americana populations are numerous. Although this study has determined a center populace and extremities for a given set of river conditions, a complete temporal and spatial study to determine the complete distributions of the copepod should be done. This would determine the extents to which the copepod may reside in differing salinities, depths, and turbidity levels at different times, conditions, and places in the river. From these data, further questions could be looked upon. These secondary projects could include mechanisms of retention, diets and related behaviors, and higher echelon food web dynamics. Vertical migration and rates of loss and gain could also be connected to the mechanisms of retention.

To go even a step further, since E. americana is dispersed throughout the Pacific and Atlantic coasts, but yet separated with respect to differing habitats, DNA studies should be considered to determine the amount of separation of populations and in turn, the mechanisms by which these differing populations may integrate. In other words, can E. americana actively move from estuary to estuary and interact on scales larger than a single estuary?
 

Acknowledgments. I'd like to express my gratitude for all those who put forth the effort, knowledge, and ideas to help complete this project. Special thanks go out to Dr. Bruce Frost, Jeffery Cordell, Byron Crump, Dean McManus, Dave Thoreson, Floyd McCroskey, and Holly Dail.
 
 

References
 

Baross, J. and B. Crump, Elevated "Microbial loop" activities in the Columbia River estuarine turbidity maximum,

In Dyer K. (ed.), Changes in fluxes to estuaries: Implications from science to management, ECSA Symposium 21, Sept. 1992, Olsen and Olsen, 1994.

Bousfield, E.L., Filteau, G., O'Neill, M. and Gentes, P., Population dynamics of zooplankton in the middle St. Lawrence estuary, In: L.E. Cronin (Editor), Estuarine Research, Academic Press, New York, NY, pp. 325-351, 1975.

Castel, J. and Veiga, Distribution and retention of the copepod Eurytemora affinis hirundoides in a turbid estuary,

Mar. Biol., 107: 119-128, 1990.

Dawson W.A. and L.J. Tilley, Measurement of salt-wedge excursion distance in the Duwamish River estuary, Seattle, Washington, by means of the dissolved-oxygen gradient, U.S. Geol. Survey Water-Supply Paper 1873-D, 1972.

Hough, A. R. and E. Naylor. Endogenous rhythms of ciratidal swimming activity in the estuarine copepod Eurytemora affinis (Poppe), J. Exp. Mar. Ecol., 161: 27-32, 1992.

METRO, Water quality assessment of the Duwamish estuary, Washington, 1983.

Miller, C. B., The zooplankton of estuaries, In: B.K Ketchum (Editor), Estuaries and Enclosed Seas, Elsevier Scientific Publishing Company, Amsterdam, pp. 103-149, 1983.

Morgan, C., Sink or swim: Copepod population maintenance in the Columbia River estuarine turbidity maxima region, Thesis, School of Fisheries, University of Washington, Seattle, WA, USA, 1993.

Morris, A. W., Mantoura, R. F. C., Bale, A. J., Howland, R. J. M., Very low salinity regions of estuaries: important sites for chemical and biological reactions. Nature, 274: 678-680, 1978.

Roddie B.D., R.J.G. Leaky and A. J. Berry, Salinity-temperature tolerance and osmoregulation in Eurytemora affinis (Poppe) (Crustacea: Calanoida) in relation to its distribution in the zooplankton of the upper reaches of the Forth estuary, J. Exp. Mar. Biol. Ecol., 79: 191-211, 1984.

Saunders, J.F., Distribution of Eurytemora affinis (Copepoda: Calanoida) in the southern Great Plains, with notes on zoogeography, J. Crust. Biol., 13(3): 564-570, 1993.

Soltanpour-Gargari, A. and Wellershaus, S., Very low salinity stretches in estuaries - the main habitat of Eurytemora affinis, a plankton copepod. Veroff. Inst. Meeresforsch. Bremerh. 20: 103-117, 1987.

United States Geological Survey, Coastline Extractor, coastline mapping software via internet, http://crusty.er.usgs.gov/coast/getcoast.htm, 1996.

Von Vaupel-Klein, J.C. and R.W. Webber, Distribution of Eurytemora affinis (Copepoda: Calanoida) in relation to salinity: field and laboratory observations, Neth. J. Sea Res., 9: 297-310, 1975.

Warner, E.J. and Fritz, R.L., The distribution and growth of Green River Chinook salmon (Ocorhynchus tshawytscha) and Chum salmon (O. keta) outmigrants in the Duwamish estuary as a function of water quality and substrate. Test report by Mukilshoot Indian Tribe, Fisheries Dept., Water Resources Div., 1995.

Williams, L.W., Notes on marine Copepoda of Rhode Island. Amer. Natural., 40: 639-660, 1906.
 
 
 
 
 
 

Sean R. Avent, School of Oceanography, University of Washington, 357940, Seattle, WA 98195-7940. (email: savent@ocean.washington.edu)
 

Present address:  San Francisco State University, Romberg-Tiburon Center for Environmental Studies and Department of Biology, 3150 Paradise Dr., P.O. Box 855, Tiburon, CA 94920.  (email: savent@sfsu.edu)