Recent Submissions

• Ecological effects of invasive European bird cherry (Prunus padus) on salmonid food webs in Anchorage, Alaska streams

Invasive species are a concern worldwide as they can displace native species, reduce biodiversity, and disrupt ecological processes. European bird cherry (Prunus padus) (EBC) is an invasive ornamental tree that is rapidly spreading and possibly displacing native trees along streams in parts of urban Alaska. The objectives of this study were to: 1) map the current distribution of EBC along two Anchorage streams, Campbell and Chester creeks, and 2) determine the effects of EBC on selected ecological processes linked to stream salmon food webs. Data from the 2009 and 2010 field seasons showed: EBC was widely distributed along Campbell and Chester creeks; EBC leaf litter in streams broke down rapidly and supported similar shredder communities to native tree species; and EBC foliage supported significantly less terrestrial invertebrate biomass relative to native deciduous tree species, and contributed significantly less terrestrial invertebrate biomass to streams compared to mixed native vegetation, but riparian EBC did not appear to affect the amount of terrestrial invertebrate prey ingested by juvenile coho salmon (Oncorhynchus kisutch). Although ecological processes did not seem to be dramatically affected by EBC presence, lowered prey abundance as measured in this study may have long-term consequences for stream-rearing fishes as EBC continues to spread over time.
• The use of aerial imagery to map in-stream physical habitat related to summer distribution of juvenile salmonids in a Southcentral Alaskan stream

Airborne remote sensing (3-band multispectral imagery) was used to assess in-stream physical habitat related to summer distributions of juvenile salmonids in a Southcentral Alaskan stream. The objectives of this study were to test the accuracy of using remote sensing spectral and spatial classification techniques to map in-stream physical habitat, and test hypotheses of spatial segregation of ranked densities of juvenile chinook salmon Oncorhynchus tschwytscha, coho salmon O. kisutch, and rainbow trout O. mykiss, related to stream order and drainage. To relate habitat measured with remote sensing to fish densities, a supervised classification technique based on spectral signature was used to classify riffles, non-riffles, vegetation, shade, gravel, and eddy drop zones, with a spatial technique used to classify large woody debris. Combining the two classification techniques resulted in an overall user's accuracy of 85%, compared to results from similar studies (11-80%). Densities of juvenile salmonids was found to be significantly different between stream orders, but not between the two major drainages. Habitat data collected along a 500-meter stream reach were used successfully to map in-stream physical habitat for six river-kilometers of a fourth-order streams. The use of relatively inexpensive aerial imagery to classify in-stream physical habitats is cost effective and repeatable for mapping over large areas, and should be considered an effective tool for fisheries and land-use managers.
• Growth of juvenile chinook salmon (Oncorhynchus tshawytscha) as an indicator of density-dependence in the Chena River

In management of Pacific salmon, it is often assumed that density-dependent factors, mediated by the physical environment during freshwater residency, regulate population size prior to smolting and outmigration. However, in years following low escapement, temperature may be setting the upper limit on growth of juvenile chinook salmon Oncorhynchus tshawytscha during the summer rearing period. Given the importance of juvenile salmon survival for the eventual adult population size, we require a greater understanding of how density-dependent and independent factors affect juvenile demography through time. In this study we tested the hypotheses that (1) juvenile chinook salmon in the Chena River are food limited, and (2) that freshwater growth of juvenile chinook salmon is positively related with marine survival. We tested the first hypotheses using an in-situ supplemental feeding experiment, and the second hypothesis by conducting a retrospective analysis on juvenile growth estimated using a bioenergetics model related to return per spawner estimates from a stock-recruit analysis. We did not find evidence of food limitation, nor evidence that marine survival is correlated with freshwater growth. However, we did find some evidence suggesting that growth during the freshwater rearing period may be limited by food availability following years when adult escapement is high.
• A total environment of change: exploring social-ecological shifts in subsistence fisheries in Noatak and Selawik, Alaska

Arctic ecosystems are undergoing rapid changes as a result of global climate change, with significant implications for the livelihoods of arctic peoples. In this thesis, I use ethnographic research methods to detail prominent environmental changes observed and experienced over the past few decades and to document the impact of these changes on subsistence fishing practices in the Inupiaq communities of Noatak and Selawik in northwestern Alaska. Using in-depth key informant interviews, participant observation, and cultural consensus analysis, I explore local knowledge and perceptions of climate change and other pronounced changes facing the communities of Noatak and Selawik. I find consistent agreement about a range of perceived environmental changes affecting subsistence fisheries in this region, including lower river water levels, decreasing abundances of particular fish species, increasingly unpredictable weather conditions, and increasing presence of beaver, which affect local waterways and fisheries. These observations of environmental changes are not perceived as isolated phenomena, but are experienced in the context of accompanying social changes that are continually reshaping rural Alaska communities and subsistence economies. Consequently, in order to properly assess and understand the impacts of climate change on the subsistence practices in arctic communities, we must also consider the total environment of change that is dramatically shaping the relationship between people, communities, and their surrounding environments.
• Foods and foraging ecology of oldsquaws (Clangula hyemalis L.) on the arctic coastal plain of Alaska

The study was conducted from June to September during 1979 and 1980 in the the West Long Lake area of the National Petroleum Reserve-Alaska. Additional oldsquaws were collected in the inland wetlands near the northwest boundary of the reserve at Ice Cape. West Long Lake and the adjacent Goose Lake are located 15 miles south of the Beaufort Sea and immediately west of Teshekpuk Lake.
• River features associated with chinook salmon spawning habitat in Southwest Alaska

Chinook salmon (Oncorhynchus tshawytscha) is a highly valued traditional, subsistence, and commercial resource in Southwest Alaska. Stream habitat availability is a major component influencing salmon productivity. The objective of this study is to identify river features associated with spawning habitat, and describe upper and lower boundaries of chinook salmon spawning on the Tuluksak River. River distances, elevation, salmon locations, spawning sites, and habitat observations were collected along 75 river kilometers of the Tuluksak River primarily within the Yukon Delta National Wildlife Refuge. Habitat and salmon observations were grouped into strata along the length of the river for comparison and analysis. Chinook salmon were observed spawning in the upper 45 river kilometers of the study area. Map-based observations of elevation and channel sinuosity correlate better with chinook salmon spawning than in stream habitat measurements along the Tuluksak River. The upper boundary of chinook salmon spawning in the Tuluksak River was outside of our study area. The lower boundary for chinook salmon spawning habitat on similar rivers might be determined by examining elevation, sinuosity, and channel features from remote images or maps prior to conducting field studies.
• Black bear denning ecology and habitat selection in interior Alaska

To identify conflicts between existing black bear (Ursus americanus) management and human activity on Tanana River Flats, Alaska, we monitored 27 radio-collared black bears from 1988-1991. We compared denning chronology, den characteristics, den-site selection, and habitat selection across sex, age, and female reproductive classes. Mean den entry was 1 October and emergence was 21 April, with females denned earlier and emerging later than males. Marshland and heath meadow habitats were avoided, and willow-alder was selected for den-sites. Eighty-three percent of dens were excavated, 100% contained nests, 18% were previously used, and 29% had flooded. Black bears selected black spruce-tamarack and birch-aspen significantly more, and marshland and heath meadow significantly less than available. Marshland and birch-aspen were used significantly more in spring than autumn. Marshland was used less than available by all bears in all seasons. Special habitat or den-site requirements are not critical for management of Tanana River Flats black bears.
• Molecular systematics and biogeography of long-tailed shrews (Insectivora: Sorex) and northern flying squirrels (Rodentia: Glaucomys)

Insight into phylogenetic and biogeographic relationships among several mammalian taxa in western North America was provided with DNA sequences of two mitochondrial genes (cytochrome b and ND4). Members of two species complexes of long-tailed shrews (genus Sorex ) and northern flying squirrels (genus Glaucomys) were examined, and a common theme of responses to past climate change and glacial cycles was evident. Diversification events indicated by the DNA sequences provide new perspectives regarding the deep and shallow history of these taxa. Analysis of seven species of the Sorex cinereus complex (and related species) revealed two major clades within the complex, Northern and Southern. These generally corroborate proposed morphological relationships and correspond to broadly defined habitat affiliations (xeric and mesic), respectively. Within the Northern clade, amphiberingian species represented a monophyletic group suggesting Beringia was a center of endemism. Next, five species of the S. vagrans complex and related species were assessed. Significant molecular variation was revealed that does not correspond to morphological differences within the complex. Two major clades within S. monticolus were observed, a widespread Continental clade (Arizona to Alaska, including S. neomexicanus) and a restricted Coastal clade (Oregon to southeast Alaska, including S. bairdi and S. pacificus). A regional examination of genetic variation in the northern flying squirrel in southeast Alaska was also performed. Results suggested that southern islands in the Alexander Archipelago were the result of recent colonization (founder event). Finally, a comparative phylogeographic analysis of a reduced data set (S. monticolus), a molecular data set for the American Pine Marten, Martes americana, and other published molecular studies were used to reexamine the role of glacial refugia in the biogeography of the north Pacific coast. Previous ideas regarding purported refugia may be overstated and may be the result of limited geographic sampling. This thesis provides new perspectives on processes (e.g., post-glacial colonization) driving mammalian phylogenetic and biogeographic structuring in western North America.
• Molecular phylogenetics of arvicoline rodents

The impetus for this dissertation was an interest in geographic variation in Microtus longicaudus with a particular focus on populations in the Alexander Archipelago of southeastern Alaska. To establish a framework for interpreting intraspecific variation in M. longicaudus, I examined the phylogenetics of 28 species of the genus Microtus, including all North American species (Chapters 2 and 4). That study, which corroborates a rapid pulse of diversification noted in the fossil record, necessitated a deeper phylogenetic perspective. Thus, a third objective of the dissertation was to investigate relationships among genera of arvicolines within the framework of other murid rodents. I examined variation in the mitochondrial cytochrome b and ND4 genes using maximum parsimony, distance, and maximum likelihood phylogenetic analyses. Relationships at several taxonomic levels appear intractable due to rapid accumulation and survival of genetic lineages. These rapid radiations were found among species, genera, and possibly subfamilies; however, strong support at these levels for other taxa (e.g., the monophyly of Microtus) suggests these genes have strong phylogenetic signal. Many of the well-supported sister species pairs within Microtus (Chapters 2 and 4) had been previously identified based on morphologic or allozyme work (e.g., M. pennsylvanicus and M. montanus, M. pinetorum and M. quasiater). The sequence data supported a clade of taiga dwelling species in North America and a clade of eastern and central Asian species. The southernmost arvicoline species of Mexico and Guatemala, though previously suggested to be derived from a single ancient invasion, did not appear to be either ancient or monophyletic. Within M. longicaudus, a large east-west phylogeographic break was detected that is equivalent in genetic distance to other sister species pairs in the genus. This break may indicate mid to late-Pleistocene differentiation (Chapter 3) within the genus. At higher latitudes, populations of M. longicaudus exhibited evidence of recent range expansion including absence of correlation between geographic and genetic structure; and pairwise mismatches among DNA sequences with a single peak and few differences.
• Production of vascular aquatic plants in wetlands of Alaska: A comparative study

I examined the effects of climate and hydrology on aboveground biomass of macrophytes in wetlands across Alaska by investigating the effects of latitude, July mean air temperature, lake type (open, periodically inundated, and closed), hydrology, and water and sediment chemistry on emergent and submersed vascular plant biomass to determine environmental variables that influenced wetland plant growth. I sampled aboveground biomass of macrophytes in four wetland complexes within Alaska: Kenai and Tetlin National Wildlife Refuges, Minto Flats State Game Refuge, and the Arctic Coastal Plain near Prudhoe Bay, Alaska. In addition to peak aboveground biomass, I also collected water and sediment samples from each lake that were analyzed for water temperature, color, alkalinity, turbidity, pH, orthophosphate, $\rm NO\sb3/NO\sb2$-N, NH$\sb4\sp+$, and total sediment C, N, and P. I found a quadratic relationship between emergent plant biomass and latitude. Minto, the second most northern site, had the greatest plant biomass, Prudhoe Bay, the most northern site had the least, and Kenai and Tetlin had moderate levels of biomass. I found a positive linear relationship between emergent plant biomass and July mean temperature, suggesting that on-site summer condition is important in predicting biomass. Submersed plant biomass was better related to alkalinity, turbidity and sediment P than to latitude, which suggests that climate is not as important in predicting submersed plant biomass as it is in predicting emergent plant biomass. Emergent plant biomass differed spatially and temporally, while submersed plant biomass showed no distinct patterns in variation across the landscape and with changes in hydrologic input. Many water and sediment chemistry variables differed among lake types and between flood regimes. Emergent plant biomass was associated with changes in water level as well as changes in water. Plant species composition differed among lake types and tended to change with flood regime as well. A separate suite of species occupied closed lakes, while open and periodically inundated lakes tended to contain more similar plant species. Both climate and hydrology appear to have a significant impact on emergent and submersed plant biomass and species composition in wetlands of Alaska. These spatial and temporal differences have direct influences on secondary producers living in wetlands of Alaska.
• Proximate and ultimate control of reproductive effort in northern shovelers (Anas clypeata) nesting at Minto Flats, Alaska

The purpose of this study was to examine factors that influence reproductive effort of female Northern Shovelers (Anas clypeata) nesting at Minto Flats, AK during summer 1991-1993. I investigated the importance of endogenous nutrient reserves to females during egg production and examined changes in organ weights and intestine lengths through the reproductive cycle. Changes in organ weights and intestine length were similar to those of shovelers nesting in Manitoba. Females used neither somatic lipid reserves, protein reserves, nor mineral reserves to produce eggs. Individual variation in somatic lipid reserves was explained by body size and nest initiation date, while variation in somatic protein reserves was explained by standardized nest initiation date. Somatic mineral variation was explained by differences among years. Neither somatic protein nor mineral reserves were reduced during incubation, but somatic lipid reserves decreased significantly. I conclude that endogenous nutrient availability does not proximately limit clutch size during laying for this population, possibly due to high productivity of interior Alaska wetlands and long days. Little is known about nest attendance behavior of ducks in the subarctic; therefore, I examined shoveler nest attendance patterns at Minto Flats to determine if observed patterns differed from those documented for shovelers nesting in Manitoba, Canada. Shovelers nesting at Minto were less attentive and took more frequent, longer recesses than shovelers in Manitoba. I examined patterns of nest attendance during incubation in relation to clutch volume and female weight loss to determine if females make tradeoffs between energy invested in the clutch and energy invested in incubation. I found no evidence of energetic tradeoffs by Shovelers nesting at Minto Flats. To determine if the trait of synchronous hatching could limit clutch size for a species of the genus Anas I measured development time and metabolic rates of Mallard (Anas platyrhynchos) eggs incubated in a constant environment. Females varied in length of time their eggs required to reach the star-pipped stage of hatch. Metabolic rate of eggs varied positively with position in the laying sequence and varied among females. These results indicate that metabolic rate may act as a synchronization mechanism for hatch.
• Population status and patterns of distribution and productivity of kittiwakes on St. George Island, Alaska

I studied populations, distributions, and reproductive performance of red-legged and black-legged kittiwakes on St. George Island in the summers of 1993-1995, where populations of both species have experienced generally poor reproductive performance and population declines of ca. 40% over the past 20 years. In 1995, I conducted a whole-island census of kittiwakes on St. George Island and found estimated breeding populations of 193,930 red-legged kittiwakes (81% of their global population), and 62,568 black-legged kittiwakes. In addition, I analyzed census trends on 51 land-based census plots on St. George Island and found that numbers of both species have stabilized in recent years. I experimentally evaluated the hypothesis that nesting red-legged kittiwakes on St. George Island are competitively displaced by larger-bodied black-legged kittiwakes to narrower rock ledges and higher elevations. I determined nest-site preferences of both species by attaching narrow and wide artificial nesting ledges within high-and low-elevation areas of St. George Island and found no evidence of competitive displacement: red-legged kittiwakes preferred narrow ledges, black-legged kittiwakes preferred wide ledges, and both species preferred ledges in areas where conspecifics nested at high density. Multiple regression analyses suggested that kittiwakes breed earlier and more successfully in summers preceded by cold winters and that inter-annual variability in kittiwake breeding success was unrelated to weather conditions during the breeding season itself. These results suggest that winter weather has indirect effects on breeding kittiwakes by influencing prey abundance several months later. Furthermore, strong winds impaired growth rates of kittiwake chicks in exposed nest sites and the growth of black-legged kittiwake chicks relative to red-legged kittiwake chicks. I identified two prominent patterns of within-colony spatial variability in kittiwake productivity and suggest that patchy "bird quality" or localized "information neighborhoods" may be responsible because traditional explanations do not apply. The "information neighborhood" is a new hypothesis which proposes that individuals are influenced by the breeding status of neighbors because their status represents an additional source of information about current breeding conditions that can be used to better tailor parental investment.
• Comparative foraging ecology and social dynamics of caribou (Rangifer tarandus)

The Southern Alaska Peninsula Caribou (Rangifer tarandus) Herd (SAPCH) and its two sub-groups were the focus of a study addressing the hypotheses: (1) food limitation during winter caused a decline in the herd; and, (2) higher calf productivity within the Caribou River group than within the Black Hill group was related to greater forage availability on the seasonal ranges of the Caribou River group. Intense, systematic range and calving surveys in 1991 and 1992 supported the hypothesis of food limitation during winter, and indicated that greater calf production in the Caribou River group was related to earlier commencement of the season of plant growth and greater forage availability on the summer range of that group, coupled with earlier parturition among females of the Caribou River herd. In a comparative study involving the two SAPCH groups and the West Greenland Caribou Herd, daily variation in sizes of foraging groups, densities of caribou within feeding sites, distances between individuals within feeding sites, distances moved by foraging groups, and frequency of group movement was modeled using the following ecological parameters: predation risk, insect harassment (by mosquitos), range patchiness, feeding-site patchiness, feeding-site area, and range-wide density of caribou. Models revealed that intraseasonal social dynamics of foraging caribou were governed in most instances by patterns of forage availability and distribution across landscapes and within feeding sites, in some instances by insect harassment and social pressures, but in no instance by levels of predation risk inherent to the ranges on which they foraged. In a study of the interrelationships between characteristics of graminoids and intensity of grazing by caribou, vegetation on each of the Black Hill and Caribou River ranges was sampled and tested for responses to clipping. Biomass density (g/m$\sp3$) of forage, shoot density (#/m$\sp2$), and nutrient and mineral densities (g/m$\sp3$) and concentrations (g/100g tissue) correlated positively with use of sites by caribou. Productivity and responses to clipping were independent of previous use, but consistent within ranges. These results indicate that caribou are sensitive to local variation in forage quantity and quality, and preferentially use sites with higher returns of nutrients and minerals.
• Application of decision analysis in the evaluation of recreational fishery management problems

Fisheries management is a decision-making process, yet typically formal decision analysis techniques are not used in structuring problems, quantifying interactions, or arriving at a prioritized solution. Decision analysis tools are applied in the decision-making process for Alaska's recreational fisheries management as a means to reduce risk in management at the policy (Chapter 2) and field (Chapter 3) levels. In Chapter 2 the analytic hierarchy process is applied to the recreational fishery for chinook salmon (Oncorhynchus tshawytscha) in the Kenai River. Model structure is developed through an iterative interview process involving individuals asked to represent the perspectives of 15 different stakeholders. Individual stakeholder judgments are combined using a geometric mean, and maximax and maximin criteria. The sensitivity of the results to under-representation is explored through various models. Despise the contentious differences of perspective represented among stakeholders, the analytic hierarchy process identifies management options that enjoy broad support and limited opposition. In Chapter 3 decision analysis is applied to the recreational spear fishery for humpback whitefish (Coregonus pidschian) in the Chatanika River. A modified form of catch-age analysis is used to combine information derived from creel surveys and run age composition with auxiliary information in the form of mark-recapture estimates of abundance. Four systems are used in weighting annual observations: prior beliefs regarding their reliability, by the inverses of their variances, through a combination of these two weighting schemes, and equal (no) weights. The perception-weighted model generates the most reasonable estimates of abundance, which are relatively precise and associated with small bias. Forecasts of mature exploitable abundance are calculated based on various recruitment scenarios, maturity schedules, and exploitation rates. From these outcomes, the odds of stock abundance occurring below a threshold level are presented. By applying decision analysis methodologies which incorporate judgments and perceptions into decision-making affecting fisheries, sensitivity to uncertain information is made explicit, components of the problem are structured, interactions among components of the problem are quantified, and options are prioritized, thus increasing the chances of finding an optimal solution.
• Ecological and physiological aspects of caribou activity and responses to aircraft overflights

I investigated the use of remote-sensing of caribou (Rangifer tarandus) activity to assess disturbance of low-altitude overflights by jet aircraft. Resource management agencies are concerned about the potential effects of these overflights on important species of ungulates. I hypothesized that low-altitude overflights would affect activity and movements of caribou, and thereby constitute a disturbance with negative consequences on energetics. I used caribou of the Delta Herd (DCH) and captive animals at the Large Animal Research Station (LARS) to address the hypotheses: caribou (1) exhibit equal activity day and night; (2) do not time activity to light; and (3) activity patterns do not change seasonally in response to daylength. Caribou were nychthemeral and exhibited uniform activity with no apparent timing to light. DCH caribou responded to seasonal changes in the environment by modifying activity (increased activity in response to insect harassment), whereas LARS caribou altered activity in response to fluctuating physiological variables (increased activity during rut). Changes in daylength did not affect activity. Data on activity from LARS and DCH caribou were compared with extant data on caribou of the Denali and Porcupine herds. Poor quality forage in winter was inferred from long resting bouts, and low availability of forage was inferred from long active bouts of post-calving caribou of the DCH. In midsummer, caribou of the DCH exhibited significantly longer active and shorter resting bouts than did LARS caribou, consistent with a moderate level of insect harassment. Responses of caribou to overflights were mild in late winter and, thus, overflights did not constitute a disturbance. Post-calving caribou responded to overflights by increasing daily activity, linear movements, incremental energy cost, and average daily metabolic rate. Energetic responses and movements were significantly related to the loudest overflight of the day. In the insect season, activity levels increased significantly in response to overflights but with no corresponding increase in linear movements or energetics. My recommendations are to prohibit aircraft overflights of caribou during calving and post-calving periods and during key feeding times in insect harassment seasons. Research indicates the possibility of more severe effects in nutritionally stressed animals.
• Demographic components of philopatry and nest-site fidelity of Pacific black brant

I investigated demographic components of nest-site fidelity and philopatry of Pacific black brant (Branta bernicla nigricans). My analyses included data I collected during summer 1990-1993, and also incorporated data obtained between 1986-1989. My studies of nest-site fidelity were limited to the Tutakoke River colony, Yukon-Kuskokwim River Delta, Alaska. Studies of philopatry and dispersal among colonies included observations at 7 breeding colonies of brant marked with tarsal tags (n = 20,147). I observed strong evidence that philopatry of brant was female biased. Probability of breeding philopatry, which was estimated with multi-state modeling techniques, was high (>0.9) and dispersal of adults among breeding colonies was rare. I developed an ad hoc estimator for natal philopatry that was unbiased by a confounding of homing, survival, and detection probabilities. Probability of natal philopatry for females was both age and density dependent. The density-dependent decline in natal philopatry may result from increased rate of permanent nonbreeding or increased probability of dispersal. Observed probability of natal philopatry for males was approximately equivalent to the relative size of their natal colony, suggesting that males pair at random with females from other colonies. Gene flow among populations of brant is largely male mediated, and I predict populations of brant will exhibit distinct mitochondrial DNAs if populations have been reproductively isolated for an adequate period of time. Probability of fidelity to previous nest sites for adults was high (>0.7). Probability of nest-site fidelity was affected by previous nesting success, age, and availability of nest sites. Phenology of nesting, nest-site selection, and clutch size of brant was affected by spring snowmelt. Dispersal of brant from traditional nest sites in years with late springs may represent a tradeoff between site fidelity and timing of nest initiation. Movement of young females from natal nest sites was a mechanism for colony expansion. I observed little evidence that site fidelity was advantageous, and concluded that quality of individual bird, environmental conditions, and demographic status may be more important determinants of breeding performance.
• Nutritional and ecological determinants of growth and reproduction in Caribou

I investigated the mechanisms by which differences in body weight and body composition (fat, protein) of female caribou (Rangifer tarandus granti) from the Central Arctic and Porcupine herds might determine changes in pregnancy rate and calf growth. Allometric relations between chemical components and body weight variables were highly significant, despite tremendous seasonal changes in composition. Between October 1989 and May 1990, body fat and body protein of adult females of the Central Arctic Herd declined by maxima of 45 and 29%, respectively; an additional 32% of fat was lost by July. Extensive mobilization of fat and protein indicates winter undernutrition. Marked hypertrophy of liver and kidneys in summer suggests the presence of mobilizable protein reserves. Birth weights of calves were similar between sexes, but male calves grew relatively faster during summer and were significantly heavier than females in autumn. Both fat content and growth rate of calves declined between 4 and 6 weeks post-calving, perhaps in response to insect harassment. Weight gains of wild calves were greatly reduced or absent after 100 d of age, while captive calves continued to grow until 175 d, suggesting that first-summer growth of caribou is determined in part by nutrient availability. Birth weight and growth rate of wild calves from birth to 3-4 weeks of age accounted for nearly 79% of the variability in autumn weights, again implying summer nutrient limitation. Female caribou were unable to entirely compensate for the metabolic and ecological costs of lactation: in autumn, lactating females had 42% less fat and 9% less protein than nonlactating females. Unlike females from the Central Arctic Herd, those from the Porcupine Herd did not demonstrate compensatory weight gains over summer; instead, autumn weight was highly correlated to June weight. Probability of pregnancy was positively correlated with body weight and fat content in early winter. Females that extended lactation into November were less fertile than predicted by body size or condition. I believe that these females were exhibiting lactational infertility.
• Seasonal diets of mink and martens: Effects of spatial and temporal changes in resource abundance

Seasonal changes in food availability and feeding habits of mink (Mustela vison) and martens (Martes americana) on Chichagof Island, Southeast Alaska, were studied through the analysis of natural abundance of stable isotopes. Dependence of the two species on marine-derived nutrients, carried to the terrestrial system via the upstream migration of spawning Pacific salmon (Onchorhynchus sp.), was investigated. Twenty-four mink and 75 martens were live-trapped repeatedly in early summer (prior to salmon runs), early autumn (post salmon runs), late winter, and in spring (during the mating season). A blood sample was obtained from each individual. In addition, 25 mink and 165 marten carcasses were obtained from trappers during late autumn 1991-1994. Concurrently, prey availability was monitored, and tissues from prey were collected. The abundance of stable isotopes in prey tissues and blood samples were compared, indicating that riverine mink depended on salmon (carcasses and fry), with little seasonal or individual variation, whereas coastal mink relied on intertidal organisms in spring and summer, but fed on salmon carcasses when they became available in autumn. In addition, analysis of blood progesterone revealed that timing of reproduction in female mink appear to be shifted, so that lactation coincided with the availability of salmon carcasses. In contrast, martens showed individual variation in their diets, with some individuals feeding exclusively on terrestrial organisms, while the diets of others include salmon carcasses. Incorporation of salmon in the diet depended largely on availability of small rodents and location of the martens home range on the landscape. Although salmon carcasses are not a preferred food item for martens, they act as a suitable alternative to maintain body condition and allow successful reproduction even in years when preferred food is not readily available.
• Prefledging survival and reproductive strategies in black brant

We develop a general model useful for estimating survival of young waterfowl between hatching and fledging. Our model allows for interchange of individuals among broods and relaxes the assumption that individuals within broods have independent survival probabilities. We consider point estimation of survival rates and corresponding variances of the point estimators. We estimated gosling survival of black brant (Branta bernicla nigricans) during summers of 1987-89 on the Yukon-Kuskokwim Delta, Alaska. Eight-two percent of females radio-marked at hatch fledged at least 1 gosling (brood success). Survival of goslings within broods was estimated by 3 methods: (1) changes in mean brood size through time, (2) observation of goslings associated with marked adults, and (3) age ratios of brant captured in banding drives. Estimates of survival within successful broods averaged 77% and ranged from 57 to 90%. Combining brood success and survival of young within broods yields estimates of overall gosling survival which averaged 64% and ranged from 77% in 1987 to 52% in 1989. We analyzed variation in egg size of black brant in relation to clutch size, laying date, female age, year, and position in the laying sequence. Egg size increased with clutch size and female age, and decreased with laying date, year, and position in the laying sequence. We did not detect a negative phenotypic correlation between clutch size and egg size. However, overlap in total clutch volumes for clutches of different sizes indicated trade offs occurred among individuals with comparable investments in their clutches. We web-tagged black brant goslings at hatch, recorded their egg size, position in the egg-laying sequence, initial brood size, hatch date, and nesting density and examined the effect of these characteristics on their probability of recapture. Larger broods from larger eggs, and with earlier hatch dates were more likely to be recaptured. There was a tendency for young females to be less successful in rearing their broods; however, this may be related to their egg size, initial brood size, and hatch date, rather than age per se.
• Influence of weather on movements and migrations of caribou

Caribou (Rangifer tarandus granti) are typified by use of calving grounds and by making twice-annual migrations between summer and winter ranges. This study used satellite technology to examine the influence of weather on calving site selection, autumn and spring movements, and timing and directionality of migrations of the Porcupine Caribou Herd (PCH) that calves in northeast Alaska and northwestern Canada adjacent to the Beaufort Sea. The reigning hypothesis that females select areas that become free of snow early for calving sites was rejected because females selected areas of $>$75% snowcover ($P=0.02$) preferentially for calving. Benefits from use of mottled snow for calving were access to vegetation in its early phenological stages and protection for their calves from predators. Access to nutritious forage and predator avoidance appeared to be the main reasons for calving site selection. Multiple linear regression models were used to examine rate and direction of autumn and spring migrations using weather data from U.S. and Canadian sources. Weather was found to be both an ultimate and an approximate influence on the rate and direction of autumn migration ($P<0.05$). Explanatory power of the equations was low ($R\sb{a}\sp2<0.41$). Proximal causes of movement were best explained by caribou tracking of vegetation phenology. Pre-rut movements in September lacked concurrence between rate and direction whereas rate and direction were related in October. Models of spring migration of parturient females indicated a common timing among years, late April and early May, and movements were significantly affected by weather ($P<0.02$), in particular snow depths and conditions that would affect foraging and traveling conditions. This study suggests that: (1) females preferentially use areas of delayed snow melt for calving, and (2) weather influences both spring and autumn migration of caribou, although the effect of weather may be more indirect than direct.