298 



DR. L. H. GOUGH. 



came from the same field, and were taken on the same day, obviously-damaged ones 

 being excluded. These bolls were divided up into 80 samples of 100 bolls each. The 

 percentage lint for the entire sample of 8,000 bolls was found at 31 "1 + - 04, the 

 standard deviation being + 0.6. The highest percentage found for any sample was 

 32*2, the lowest 29-5. The total weight of the entire sample was 17,782 grammes, 

 the average weight of the sub-samples being 222 '3 grammes. 



With a standard deviation of + - 6 on samples weighing over 200 grammes, and a 

 range of 2"7 per cent, between the lowest and the highest sample, it is obviously not 

 safe to quote percentages obtained from such small samples to beyond the nearest 

 unit, unless, of course, the sample is at the same time the entire quantity whose 

 percentage it is desired to estimate. 



Contrary to Mr. Balls' practice (" The Development of Eaw Cotton," p. 181) all 

 the samples used in this paper for ginning output were ginned first and weighed 

 afterwards. The gin used was a four inch one, and was fitted with covers to prevent 

 the loss of seeds. The samples were not weighed until, on an average, over two weeks 

 after ginning. It is believed that part at least of the reduction in weight said to be 

 observed after ginning is not due to loss of seeds, but rather to loss of moisture from 

 the seeds and more especially from the lint, as considerable heat is generated during 

 the ginning process. The delay of two weeks, mentioned above, is more than ample 

 to restore seed and lint to their normal conditions of humidity for the temperature 

 and atmospheric humidity of the moment of examination. 



Lint-weight is to a great extent influenced by the atmospheric humidity, the 

 fluctuations taking place very quickly. Figure 3 shows two graphs, one made by a 

 self-recording hygrometer, the other by the weight of one gramme of lint, under the 

 same experimental conditions as in the weight record for seed. It will be seen that 

 in this case the record traced by the light ray on the sensitive paper is in perfect 

 agreement with the hygrometer record, and is synchronous in all its changes. 



X 



II 



» 



» 



j 



9 10 M T 



> 



I 



s e ioXll2 



» 6 



3 10 M T 2 



* 6 



3 10 XII 



I 



% 



5 



3I0M T 



» 



I 



j 



BIOX 











































































































































































































































































































70 







































































































































































































































































































60 



* 



s- 





























































































- 















































































































































































































































































































































































































































































































40 

























































































































































































































































































































































































































































Fig. 3. Coincident fluctuations in humidity (solid line), and lint 

 weight (broken line) ; increasing weight is indicated by a rising 



curve. 



With seed- weight, as shown in figures 1 and 2, the influence of atmospheric humidity 

 is felt, but is subject to a lag of a few hours, whereas in the case of lint-weight it 

 corresponds simultaneously with the movements of the hygrograph. It is therefore 

 obvious that the percentage lint will also constantly be changing, especially on 

 account of the lag. That this is actually the case can be seen from figure 4, which 

 was obtained in the same way as figures 1, 2 and 3, but with 2 grammes of seed on 



