Results presented here are based on analysis of data collected and published by other researchers.  Comparison among grizzly populations and among years within the Yellowstone grizzly population reveals that essentially all parameters of reproduction and recruitment (survivorship minus net emigration) were positively correlated with indices for mean individual nutrient-energy balance and for food supply.  Effects of nutrient-energy balance exerted during gestation and infancy of a cohort continued to affects its recruitment rate to adulthood.  These results confirm and extend findings by earlier investigators, particularly Rausch (1961), Rogers (1976, 1977), and Eiler (1981).  Nutrient-energy balance was largely determined by food supply per unit bear-mass, which was a function of food supply, population density, and population infrastructure.

    Although densities of cub litters and of cubs would be positively correlated with density of adults when adults are scarce, all available data is for cases where densities of adults (especially males) were so high that the correlations were negative--indicating strongly negative density dependence, partly as a consequence of food competition.  Adult males tend to dominate in food competition as a consequence of their larger body size and aggressiveness.  The greater the abundance of adult males (in excess of the number needed for siring offspring), the less food remains for females and immatures, and thus the lower the rate of reproduction by adult females and the rates of recruitment by immatures.  Direct aggression by adult males against adult females and immatures also impaired reproduction and recruitment.

    The forms of density dependence exhibited by Yellowstone grizzlies are exceedingly complex.  When density of adult males was high, few cubs were produced, mainly daughters; these cohorts had low rates of recruitment to adulthood, even if density of males was low when the attrition occured.  By contrast, when adult males were scarce, many cub litters and cubs were born, mostly sons; these cohorts had high rates of recruitment to adulthood, regardless of density of adult males during their maturation.  There were also positive correlations between attrition of recently weaned subadults vs. concurrent density of adult males.  These results extend findings by earlier investigators, particularly Kemp (1972, 1976), Shaffer (1978, 1983), Stringham (1980, 1983), Young & Ruff (1981), and McCullough (1981).

     My results are interpreted in terms of competitive reproductive strategies of adult males vs. females.  (1) Aggression by adult males against immatures might benefit the males through  (a) nutritive value of immatures which are eaten,  (b) reducing current and future competition from victims for resources or genetic representation, or  (c) increasing opportunities to mate with mothers of the immatures.  (2) Females might minimize wastage of investment in offspring likely to be killed or prematurely exiled by the adult males.  This could explain the negative correlations between  (a)  number, sex ratio, and recruitment rate (manifesting investment) per offspring vs.  (b) adult male abundance and adult sex ratio.  (3) Selection pressures addressed by the Fisher (1930) and Trivers-Willard (1973) hypotheses, respectively, may also favor reciprocal adult-offspring sex ratios and corresponding levels of investment per cohort.

    Presentation of findings on grizzlies is accompanied by review and analysis of comparative information on other Ursidae, especially black and polar bears.  An attempt is made to integrate the bulk of current knowledge on factors governing dynamics of bear populations in order to more clearly reveal its implications for theory and management, and to facilitate development of theoretical stock-recruitment and population   models--models in which dynamics are controlled by food supply, densities of adult males and females, adult sex ratio, and age.  Although these descriptive statistical results should be most applicable to grizzly populations where bears frequently aggregate in large numbers at food concentrations, basic features of the theoretical models should be applicable to bear populations in general and perhaps to other taxa.