Overview of Results

Following this introductory chapter, I first report (Chapter II) on a comparison of the albino, Hsd:ICR, random-bred strain of house mice with wild mice from a Wisconsin population (now in press as Dohm et al. 1994: see also Richardson et al. 1994; Garland et al. in review). These comparisons sought to address how appropriate the ICR mice may be for drawing conclusions about the evolution of metabolism and locomotor performance in small rodents. The ICR strain is then used as a model system to address the extent to which individual variation for such traits as maximal sprint running speed and stamina (chapter III), BMR and exercise-induced VO₂max (chapter IV) is genetically based, using the analytical techniques of quantitative genetics (Falconer 1989; Neale and Cardon 1992; Boake 1994). We used a breeding design involving parents and their offspring, half-sibs, and full-sibs, with cross-fostering, to obtain estimates of four components of phenotypic variance and covariance (assuming no epistatic genetic effects): additive genetic effects, dominance genetic effects (confounded by pre-natal maternal effects, if present), post-natal maternal and/or common environmental effects, and specific environmental effects. The effects of variation in body mass, sex, age, measurement block, and other relevant covariates were removed by computing residuals from ordinary least squares multiple regression equations prior to genetic analyses using restricted maximum likelihood methods (Shaw 1987, 1991). In the final chapter (chapter V), I investigated the extent to which sub-organismal traits, thought to be casually linked to performance variation at the whole-animal level, can actually predict individual phenotypic variation at the whole-animal level. Appendix I presents the derivation of the linear model and likelihood function used in chapters III and IV. Perhaps the most important contribution of my dissertation work is the estimation of genetic correlations between speed and endurance and between BMR and VO₂max as partial tests of two long-standing evolutionary hypotheses concerning physiological variation. The following paragraphs briefly outline the results of chapters II-V.

Chapter II.

We conducted a common garden experiment to compare laboratory and wild house mice (trapped from a local Wisconsin horse barn) and their reciprocal hybrids for aspects of locomotor performance and metabolism (Chapter II). Only females were compared; males were used for a different study (Garland et al. in review). Analysis of covariance (using, for example, body mass and age at testing as covariates) indicated that wild and hybrid mice had significantly faster (+50%) forced sprint speeds than did laboratory mice. Wild and hybrid mice also had significantly higher VO₂max during forced exercise and greater relative ventricle masses than did the laboratory mice (see also Taylor et al. 1981). In addition, wild and hybrid mice together tended to have lower BMR and higher regulatory nonshivering thermogenesis (Richardson et al. 1994). Wild and hybrid mice also showed statistically higher swimming endurance times relative to body mass than did lab mice, although these differences were not significant when body mass was omitted as a covariate. We found no differences for relative gastrocnemius muscle mass, liver mass, hematocrit, or blood hemoglobin content. During a 7-day voluntary wheel-running test, both wild and hybrid mice ran significantly more total revolutions and at higher average velocities than did lab mice; however, the amount of time active on the wheels did not differ significantly among groups. In general, the wild and laboratory mice differed more for behavioral and/or whole-animal performance traits than for lower-level traits. Therefore, although substantial, genetically-based differentiation has occurred during the domestication of house mice, the magnitude of any differentiation is not uniform across the phenotype. The generality of our comparisons is unknown, because they involved only one strain of laboratory house mice and only a single population of wild house mice (cf. Garland and Adolph 1994). However, results from many other studies also indicate differences between domesticated and wild house mice (e.g., Wolfe 1969; Smith 1978a,b, 1985; Smith and Connor 1978; Singleton and Hay 1982). This work was published in American Journal of Physiology (1994).

Chapter III.

Both interspecific comparative data and functional studies of locomotor physiology suggest that vertebrate animals cannot both be fast and possess great stamina. We estimated the additive genetic correlation between maximal sprint running speed and swimming endurance in approximately 300 ICR house mice (both parents and their offspring) as a partial test of this trade-off hypothesis (Chapter III). Residual sprint speed and swimming endurance (corrected for appropriate covariates) showed no phenotypic correlation, nor was speed or stamina significantly correlated with body mass, either phenotypically or genetically. The narrow-sense heritability of maximal swimming endurance was statistically different from zero (h 2 = 0.3, P < 0.001); only sprint speed on the first trial day exhibited appreciable amounts of heritability (h 2 = 0.17, P < 0.1). The additive genetic correlations between sprint speed on the first trial day and all four measures of swimming endurance were large and statistically different from zero at the P < 0.1 level. These results provide partial support for the hypothesis of a necessary trade-off between these two aspects of locomotor abilities. This work was published in Evolution (1996).

Chapter IV.

Maximal (VO₂max) and basal (BMR) rates of oxygen consumption are thought to be functionally coupled in vertebrates because they share the cardiovascular and other physiological pathways. As a partial test of this hypothesis, we estimated the additive genetic correlation and the narrow-sense heritabilities of VO₂max and BMR in the ICR mice (Chapter IV). After removing the effects of body mass and other significant covariates, residual BMR and VO₂max both showed low heritability (h 2 < 0.1). Residual BMR and VO₂max were not correlated at the phenotypic or genetic levels. This lack of genetic coupling is inconsistent with the "aerobic capacity model" for the evolution of endothermy (review in Hayes and Garland 1995). This work is now in review at Genetics (2001).

Chapter V.

In the final chapter (V), I tested for significant statistical phenotypic associations between the whole-animal traits (forced maximal sprint running speed, swimming endurance, maximal and basal rates of oxygen consumption) and a series of morphological and physiological traits measured on approximately 300 male and female ICR mice. Neither speed nor endurance was related to body mass, but metabolic rates and organ masses (ventricles, gastrocnemius muscle, liver) were, as expected, positively correlated with mass. Effects of various covariates (e.g., body mass, age at measurement) and main effects (e.g., sex, measurement block) were removed statistically by computing residuals from least squares multiple regression equations. Inter-correlations between the organismal and lower-level traits were generally quite low for the residuals. This lack of correlation suggests that differences among individuals for sprint speed, swimming endurance, and maximal oxygen consumption were either largely "motivational" in origin or related to other morphological or physiological characters not included in this study. This work, plus the genetics of suborganismal traits (heart mass, liver mass, gastrocnemius muscle mass, hematocrit, hemoglobin, total cholesterol, triiodothyronine [T₃], thyroxine [T₄], and tail length) in ICR mice, is now in review at Zoology (2001).

Previous: Origin and Maintenance of the Hsd:ICR Random-bred Strain

Next: Literature Cited

Mike's Homepage

Copyright © 2006 By Bryan Johnson. All rights reserved.
e-mail the webmaster