190 

 wildlife shot by hunters (see Chapter 3 for additional information). Voucher specimens were identified 

 by Mexican biologists and deposited at the Museo de Vertebrados, Centre de Investigaciones de 

 Quintana Roo, Chetumal, Quintana Roo, Mexico. 



Statistical Methods 



The results presented in this chapter were based on three data sets: One, during 1989 and 

 1990, gardeners were interviewed to determine number of gardeners, number of gardens, annual 

 horticultural cycle, years of previous use, garden size, garden distance from X-Hazil Sur, number of 

 crops planted, number of crops per garden, and reported crop predation. Two, during 1990, seven 

 gardens were selected for an intensive study of crop phenology, density and mortality of crops, and 

 actual crop predation. Three, during 1990, track slicks at 18 sites were used to compare wildlife use 

 of gardens versus use of adjacent forest, wildlife use of the three vegetation types, and monthly 

 variation in wildlife use of the three vegetation types. 



Parametric and nonparametric tests were used to analyze garden and wildlife data. For the 

 chi-square comparison (x^) of garden aops between years, three crops (chaya, lentils, and other) were 

 combined to obtain an expected cell frequency S 5. The nonparametric Kruskal-Wallis test (x^ 

 approximation) was used for variables where the observations likely were not independent or normally 

 distributed. For example, several gardeners reported that they always planted their fields near those of 

 other family members. Several gardeners also reported that they always planted a single crop of com 

 in large gardens and mixed crops in smaller gardens. The results are presented as x^ = chi-square; x^ 

 approximation = chi-square approximation; J = mean; SE = standard error; n = sample size; d.f. = 

 degrees of freedom; and P = significance level. All statistical analyses were conducted using SAS 

 (SAS Institute Inc., 1988). 



