34 



PROCEEDINGS OF THE NATIONAL MUSEUM 



assumption that the presence of a lizard on a tree did not affect tlie 

 probability of another lizard being found on the same tree. These 

 distributions were then compared with the observed distributions 

 (table 5) and the probability that the differences were due to chance 

 were calcidated, using an X^ test. These results show that there are 

 significantly more trees with only a single male (at the .01 level) 

 than would be expected if they were distributed randomly, and 

 fewer trees with two and with no males. 



The observed distribution of the females differed from the expected 

 in the same direction (table 5) but not enough to be statistically 

 significant (at the .05 level), perhaps because of the smaller sample 

 size. 



Table 6 shows the relationship between adult males and adidt 

 females. Trees with no adidts and trees with an adult of each sex 

 occur more frequently than M^ould be expected if the distribution 

 of the sexes was independent, while trees with a single male or a 

 single female occur less frequently. 



Table 5. — Results of a count of the number of A. lineatopus on a line of trees com- 

 paring the number of adults per tree with the expected distribution if the adults 

 are distributed independently of one another (the expected distribution is cal- 

 culated from the binomial tlaeorem; the sexes are treated separately; the 

 distribution of the adult males differs from the expected at the .01 level, that 

 of the females does not) 



Table 6. — Distribution of males and females with respect to the opposite sex and 

 the null hypothesis (in parentheses) that their distributions are not related 



