Variation in the length of the eighth primary was found to occur, 

 and this primary had not completed its growth in some of the birds 

 examined. To compare the variation due to incomplete growth 

 with the variation due to individual differences, primaries seven 

 and eight, from the 23 wings, were compared. 



Table 2 and Figure 11 summarize the analyzed data from all the 

 wings that were examined, both in the field and in the laboratory. 

 In the table the data headed "primary 7" and "primary 8" are from 

 the 23 wings examined in the laboratory, and the other data are 

 presented by classes of age and sex. The presentation in Figure 11 

 is by "Dice squares" which include the range, mean, one standard 

 deviation either side of the mean and two standard errors either 

 side of the mean. Two standard errors either side of the mean 

 represent the 0.9546 confidence limits that may be placed on a 

 given sample. Thus, where these do not overlap, the difference 

 between the means is significant at that level. The coefficient of 

 variation is a measure of the variability — the smaller the coeffi- 

 cient the less the variabihty. 



In the 23 wings, the seventh primary is less variable than is the 

 eighth, V = 3.77 as compared to 5.64. This difference is caused 

 by the shortness of the eighth primary in some birds; this short- 

 ness is indicated by the longer range below the mean in Figure 19. 

 The pulpy bases of the shafts indicate that these feathers are not 

 completely grown. Contrasted with the data from primary number 

 eight, those for number seven are completely symmetrical (see 

 Figure 11) indicating a normal distribution of primary lengths. 



In samples from the field, the lengths of some feathers stood 

 apart from those of others on the lower end of the scale, indicating 

 incomplete growth in those feathers. It seemed reasonable that 

 the data might be refined by dropping these measurements from 

 consideration when comparing lengths of primaries of males and 

 females. This was done, and these categories are indicated as 

 selected males and selected females. It is apparent (see Figure 

 11) that the average length of primary eight is different between 

 males and females, and that this difference is statistically significant 

 at a high level of confidence, but that the range in length is too 

 great to permit the separation of sexes by this means, even when 

 the lower extremes (incompletely grown feathers) are eliminated 

 as in the instance of the selected groups. The average length of 

 primary number eight is not significantly different between birds- 

 of-the-year and adults in the samples available to me and it is 

 thought that a larger sample would reveal no significant difference. 



[40] 



