Data sources and their problems

There is a vast literature on body weights among adult primates, and the selection of a species-typical mass is fraught with problems (see Smith and Jungers, 1997). For a small sample such as that used here, variation of the order of a few hundred grams is probably relatively unimportant. The term 'weight' is used as synonymous with mass, because for terrestrial species there is unlikely to be discordance between the measured weight and the assumed mass. In the selection of the original data base (Lee et al., 1991), priority was given to populations for which as much information as possible was available. This will, indeed, result in some skew in the data, as many growth data are derived from captive sources, which tend to be heavier and develop more rapidly than wild populations (Leigh, 1994b). Mass data are presented for adult females, neonates and weanlings. Weaning weights were calculated from growth curves, and taken as the weight for age at the end of the major period of lactation.

Lactation duration was defined as the population-specific (where weights were derived from a population) interbirth interval - gestation time. This measure incorporates an average duration of lactational anovulation, when the majority of the infant's nutrition is derived from milk, plus some cycling time. As primates frequently lactate into the next pregnancy, defining weaning as the period of major nutritional input that affects the mother's subsequent reproduction should standardise the term across species, and thus reflect a process rather then an endpoint (Lee, 1996). Data on gestation length are 'species-average' because variation due to infant sex (Clutton-Brock, Albon and Guinness, 1989) or ecology (Silk, 1986) is in the order of days rather than months. Other data presented as species averages are adult brain mass and neonate brain mass. These have been taken from Harvey et al. (1987). Data on M1 eruption age were taken from Smith et al. (1994), and used in these analyses because previous work (reviewed by Smith, 1992) suggests M1 eruption age is an excellent life history measure.

Two 'measures' of environmental risk were incorporated in the analyses. The first was that defined by Ross (1988) in relation to predictability of the environment. Ross' measure assessed resource type, its productivity and seasonality, without relying on specific dietary variables (see also Chapter 4). Diet may be related to evolved digestive capacities, and thus more subject to phylogenetic error than are general habitat parameters. The second measure categorises potential mortality risk due to predation pressure, and is derived from a population-level assessment of predator presence, contacts between primates and predators, antipredator behaviour and observed predations (see Hill and Lee, 1998). These qualitative categories were used in preference to more quantitative vari-ables, as a means of discriminating between gross habitat qualities.

Data were limited to haplorhine species bearing single young, in order to eliminate any possible 'grade' effects or biases from including strepsirhines, and to reduce the variation introduced in individual growth for a litter size greater than one - factors recently addressed in detail by Garber and Leigh (1997). Only a small number of species is represented (n = 41) and data used are presented in the appendix to this chapter.

All data were natural log transformed to facilitate comparisons between variables and to normalise data for parametric tests. This was done in order to use multiple regression statistics, and to compute residuals from observed relationships. Residual analysis has been widely used in life history studies as a means of 'removing' any allometric effects and thus to explore variation remaining after controlling for autocorrelation. Here, multiple regressions are also used to assess autocorrelation between size variables. As such, the results are presented as least squares regression slopes, which may be less accurate and produce a lower slope. It should be noted that when comparisons are made between slopes and constants, at least some of the differences may be simply a result of the two techniques for calculating the slopes rather than demonstrating any biological difference.

As noted above, much of the analysis was conducted at the level of the species, without phylogenetic correction techniques. There are two reasons for this: firstly, to explore the variation within closely related taxa (see below); and, secondly, because many allometric studies find no effect of phylogeny on the overall variation (Martin, 1996). Rogers and Cashdan (1997) note that when closely related species experience different rates of selection, phylogenetic subtraction techniques may statistically obscure species trends.

Losing Weight Without Starving

Losing Weight Without Starving

Tired of Trying To Loose Weight And It Never Works or You Have To Starve Yourself Well Here's A Weight Loss Plan That takes Care of Your Weight Problem And You Can Still Eat. In This Book, You’ll Learn How To Lose Weight And Not Feel Hungry! In An Easy Step-By-Step Process That Enables You To Feel Good About Loosing Weight As Well As Feeling Good Because Your Stomach Is Still Full.

Get My Free Ebook

Post a comment