52 



BULLETIN 1134, U. S. DEPARTMENT OF AGRICULTURE. 



sidering only the naturally pollinated flowers, the mean numbers of 

 seeds per boll and per 100 flowers were significantly greater during 

 the second half of August, 1920, than during the first half. Whether 

 the difference, which was much more pronounced at Phoenix than at 

 Sacaton, was caused by more favorable weather or by an increase 

 in the number of pollinating insects is uncertain. 



The bolls upon which were computed the mean numbers of seeds 

 given in Table 30, with the single exception noted, were taken at 

 random, without reference to the number of locks. It has been 

 determined by a number of counts which have given practically the 

 same result, that at Sacaton the proportion of 4-locked bolls in Pima 

 cotton slightly exceeds 5 per cent, 30 practically all the others being 



2S 



20 



/S 



/O 



S 

















_1 



>— < 



i 





























/ 



/ 





\ 



\ i 

 \ i 



I 



\ 

 I 

 % 



1 

























/ 







W 

 l\ 



; \ 

 i 



\ 



\ 



\ 



\ 



\ 



\ 



\ 



/ 

 / 



/ 



\ 



\ 



\ 





















i 

 1 



I 

 t 

 i 

 I 



K 



i 



1 



HI 



i 



\ i 



I 



i 



> 



y 



\ 







V 







/ J 





i 

 t 



1 



t 





\ 



i \ 

 i > 



\ 



















, 









"7- 



-/-(J 



C/f-* 



// /2 /3 /4 /S /e /7 /S /S 20 2/ 22 23 2f 25 26 27 28 

 7/(jM8f/? or Jttas Pftf SOIZ. 



Fig. 4. — Frequency distributions of the number of seeds in one hundred 3-locked and 4- 

 locked bolls on well-grown plants of Pima cotton at Sacaton, Ariz., in 1921. The 

 bolls were collected in pairs, a 3-locked and a 4-locked boll from each plant. The 

 mean numbers of seeds were 18.6 ±0.13 for the 3-locked bolls and 21.7 ±0.19 for the 

 4-locked bolls. The frequency distribution for the 3-locked bolls, shown by the solid 

 line, is much more regular than that for the 4-locked bolls, shown by the dotted line. 



3-locked, the number of bolls having two and five locks being negli- 

 gible. It was found at Sacaton in 1921 by counts on 100 bolls of each 

 lock number taken from as many plants that the average number 

 of seeds per boll was 18.6zb0.13 for the 3-locked bolls, and 21.7±0.19 

 for the 4-locked bolls. 81 The frequency distributions for the number 

 of seeds in the 3-locked and in the 4-locked bolls are shown in 

 Figure 4. 



80 Much higher percentages have been recorded at other localities. Counts made by 

 C. G. Marshall and W. B. Camp on 40 plants in a field of Pima cotton near Bakersfleld, 

 Calif., in 1917, showed that in a total of 2,486 bolls 21.3 per cent were 4-locked. O. F. 

 Cook in 1920 counted bolls on 5 plants taken at random in a field near Porterville, Calif., 

 and found that in a total of 62 bolls, 20 (or 32.3 per cent) were 4-locked. 



81 A greater difference in sea-island cotton in the mean numbers of seeds in 3-locked 

 and 4-locked bolls is indicated by data given by Harland (21, Table II, p. 152), which 

 show that the means for 3-locked and 4-locked bolls were 17.4 and 23.2, respectively. 



