Women's fishing and women's knowledge of marine ecosystems has been given scant attention in the literature. The few studies that do exist indicate that women's fishing provides a consistent and stable form of protein for subsistence. In the early part of this century, women's net fishing on Kosrae was a highly developed and varied activity that included specialized gear and techniques. Women's fishing remains an important activity on Kosrae although only a few individuals still practice traditional fishing methods. This paper reviews past and present forms of women's fishing practices on Kosrae and examines the social and technological changes in women's fishing over the last century.

We evaluated four techniques for estimating densities of the Brown Tree Snake, Boiga irregularis, which is nocturnal and difficult to detect in natural habitats: snake traps, time-constrained visual censuses, distance-constrained visual censuses, and mark-recapture programs. Snake traps are best suited to small areas(<5 ha) and can be maintained by individuals without special skills; but they are labor intensive and have an extremely high sampling coefficient of variability (65-200%). Furthermore, traps do not lend themselves to either relative or absolute density comparisons between sites. Both types of visual censuses provide estimates of relative abundance only, have a high coefficient of variability (40-90%), and are sensitive to differences in snake behavior among sites, uncontrolled variability in environmental conditions, observer skill, and habitat differences. A 95% confidence of detecting a 25% difference between the mean values of two sites requires about 100-200 visual censuses for each site. Mark-recapture provides absolute population estimates, but the method can only be applied to snake populations with a high density (>20/ha). Mark-recapture requires an intensive commitment to capturing many snakes in a short time and a high density trail network or other access to all parts of a study area.

A preliminary checklist of 314 taxa of benthic marine algae is provided for the Fiji Islands, South Pacific, comprising 11 Cyanophyceae, 99 Chlorophyceae, 36 Phaeophyceae, and 168 Rhodophyceae. Included are all previously published records, with the systematic arrangement and nomenclature brought up to date. The flora is relatively poorly known, and many areas have yet to be phycologically studied, such as Rotuma, much of the Lau Group, most of Vanua Levu and Kandavu.

Twenty-one species of birds were recovered from preliminary excavations in 1990 at a late Holocene bone deposit in Payapai Cave, Rota, Commonwealth of the Northern Mariana Islands. Twelve of these species are extinct or extirpated on Rota. Three of these (Audubon's Shearwater, Puffinus lherminieri; Micronesian Megapode, Megapodius laperouse; Island Swiftlet, Collocalia vanikorensis) are known from Rota in the 19th or 20th century. Two others (White-browed Crake, Poliolimnas cinereus; Guam flycatcher, Myiagra cf. freycineti) are previously unknown on Rota but have been recorded on Guam. Three species (Blue-gray Noddy, Procelsterna cerulea; swamphen?, Porphyrio? sp.; Micronesian Pigeon, Ducula oceanica) have not been recorded previously from the Mariana Islands but still occur elsewhere in Micronesia. Four species are extinct and undescribed (a small, perhaps flightless duck, Anatidae new sp.; giant ground-dove, Gallicolumba new sp.; large parrot, Psittacidae new sp.; large parrotfinch, Erythrura new sp.). An unnamed cave on Rota yielded a single bone of the Common Moorhen, Gallinula chloropus, also extirpated on Rota. Although more bones are needed to refine certain identifications and describe new species, the preliminary results from Rota indicate that great losses of birds occurred in the Mariana Islands during the past few millennia. Probably the same is true throughout Micronesia, where the prehistoric record of birds, confined thus far to Rota, is much more limited than in Polynesia.

A remnant native limestone forest in the Marpi region, Saipan, was studied to characterize the species composition, relative species importance, tree basal area, and tree density in that habitat. Using the point-quarter sampling technique on 400 canopy and understory trees each, I found that mean canopy and understory basal area were, respectively, 45140 ± 24144 and 1460 ± 749 m²/ha. Mean trees/ha in the canopy and understory were, respectively, 794 ± 174 and 2579 ± 770. Pisonia grandis overwhelmingly dominated in canopy basal area, whereas Cynometra ramiflora, Pouteria obovata, and Guamia mariannae dominated in the understory. Cynometra ramiflora occurred most densely in the canopy, and C. ramiflora and G. mariannae had the highest densities in the understory. The relative measure of forest prevalence, the importance value, showed that P. grandis and C. ramiflora were the principal canopy species, and C. ramiflora and G. mariannae were the major understory species. A model of forest dynamics is proposed in which P. grandis and other species largely absent from the understory may be maintained in the forest through disturbance, which might facilitate vegetative propagation, flowering, or survival of seedlings. Commonly fruiting, shade tolerant species like C. ramiflora, which were frequent in both the canopy and understory, may come to predominate under conditions of low forest disturbance. Despite centuries of human-related disruption, forest composition remains overwhelmingly native, suggesting that native forest is resistant to invasion by alien species.

The labrid genus Cirrhilabrus is represented by seven species at islands in the northwestern Pacific which include southern Japan, Taiwan, and the Marianas. Five are named species: cyanopleura, exquisitus, lanceolatus, lunatus, melanomarginatus, and temminckii. Two are described as new: katherinae from the Izu and Mariana Islands is characterized by 5 median predorsal scales, 21-23 lateral-line scales, 15-18 gill rakers, body depth 3.4-3.7 in SL, rounded to slightly rhomboid caudal fin, long pelvic fins in the male, small size (largest, 71 mm SL), and a color pattern of red dorsally, abruptly white ventrally, with a middle black band in the dorsal and anal fins and a broad red band distal to it; and rubrimarginatus which ranges in the western Pacific from Okinawa south to Fiji and Tonga; it has 5 median predorsal scales, 21-25 lateral-line scales, 17-20 gill rakers, body depth 2.85-3.1 in SL, rounded caudal fin, long pelvic fins in the male, large size (largest, 121.5 mm SL), and a color pattern for which the most distinctive feature is a broad red band in outer part of the dorsal and caudal fins.

Jim Comber has for many years explored the interior of Java with his keen eyes and camera, always on the lookout for orchids. That may not be unique, but he has summarized here in this handsome volume much of his own experience, basing it solidly on the nearly two centuries of botanical exploration in the island.

(This is the first paragraph. There is no abstract.)

Perhaps because of their distance from other islands, the Marianas have a relatively high proportion of ferns in the vascular flora. Fern spores are light and travel easily on the wind. On the other hand, the number of orchids is small compared to islands closer to the mainland (see other book reviews this issue). Ninety species of ferns, and related plants such as Psilotum. Selaginella, and Lycopodium, are described and beautifully illustrated in color in this book. Thirty species of orchids, not including non-naturalized introductions, are also included. Of these orchids, four are endemic to the Marianas and seven more occur only in Micronesia. Not surprisingly, there are few endemic ferns; one, Thelypteris gretheri, commemorates David Grether, co-author with W.H. Wagner of the first book on ferns of Guam (Pteridophyta of Guam, Bishop Museum, 1948). Each species is shown in one or two clear photos; detail photos of the ferns often show the arrangement of the reproductive areas (sori), and of course orchid details show the flowers.