The southernmost extension of winter ice cover varies interannually and on longer time scales, reflecting large‐scale changes in driving forces, especially in the position and intensity of the winter Aleutian Low Pressure System. A conspicuous pattern is alternating warm and cool periods of several years' duration. These variations in sea ice cover are reflected in the character of a subsurface cold pool, formed as stratification isolates the deeper cold waters from surface exchanges. The cold pool is better developed and more extensive in summers that follow deep southward penetration of winter sea ice. Interannual and decadal‐scale variations in the distributions of some fish stocks reflect those of ice and thermal conditions. In particular, the distribution of walleye pollock, Theragra chalcogramma, varies significantly with multiannual cool and warm years while Arctic cod, Boreogadus saida, is only present within the cold pool. The relation among climate variations, sea ice cover, subsurface thermal conditions, and fish distribution provides information on how climate affects marine ecosystems and may also have practical application in predicting fish distributions.

During the late summer of 1997, for the first time ever recorded, most of the continental shelf of the eastern Bering Sea was covered by aquamarine waters (Figure 1), resulting from a massive bloom of coccolithophores (Figure 2). Light reflecting off the calcium carbonate plates of the flagellated coccolithophores gave the water its anomalous color, which was first observed in July. The bloom was also clearly visible from space, as shown by some of the first images from the multispectral sea‐viewing wide‐field‐of‐ view sensor (SeaWiFS) scanner in September. Light penetration into the water column, essential for primary production by diatoms and other phytoplankton, was markedly reduced. This shift potentially altered the trophic dynamics throughout the food web of one of world's most productive ecosystems.

In 1997 and 1998 seeds were collected from the seagrass, Zostera marina L. (Zosteraceae) from Long Island, New York, USA (var. typica), Shaw Island, Washington, USA (var. latifolia) and Izembek Lagoon, Alaska, USA (var. izembekensis) to determine if seed size varied between the varieties, if seed size varied within the same variety, and if the relationship between seed length and seed weight was linear. Measurements of seed length revealed statistically significant differences in this metric between varieties associated with smaller intertidal plants (var. typica and var. izembekensis) and larger plants from the subtidal region (var. latifolia). It was also found that seed length varied within the same variety and verified that the cross‐sectional shape of these seeds was primarily elliptical. Using seed cross‐sectional area and weight from a randomly selected batch of seeds from the three locations sampled in 1998, a linear model was created to determine seed weight and seeds from var. latifolia were found to be heavier than seeds from var. typica. The potential advantage of heavy seeds for subtidal seeds in search of safe sites is discussed, together with the merit of using the model to calculate the weight of other elliptical seeds within the Zosteraceae and other seagrass families. © 2003 The Linnean Society of London, Botanical Journal of the Linnean Society, 2003, 142, 281–288.