172 THE DEVELOPMENT OF RAW COTTON 



ordinates, and very little supervision was required by 

 the author or his assistant, since every set of plants 

 recorded constituted a check on the exactness of the 

 records from other sets. If the curves presenting the 

 data from two identical sets of plots did not coincide 

 after computation each day, slackness on the part of the 

 Observers was suspected at once, and it is only fair to 

 say that such occurrences were very rare. The Observers 

 could not possibly concoct their results, since they had 

 only a vague notion of the nature of the dozens of different 

 groups of plants which they were observing, and the 

 actual numerical records meant nothing even to us, until 

 they had been grouped, totalled, divided, and plotted on 

 squared paper. Our records of crop condition at Giza 

 in 1913 attained nearly to absolute precision day by day 

 from May to November. 



All data were taken from definite groups of plants 

 called "observation rows," containing a known number 

 of individuals, and the results were all computed down to 

 terms of an average plant. Presentation of the con- 

 dition of such a plant as should be the average for a whole 

 field, or even larger areas, was obtained by sampling- 

 various portions with a regular " scatter " of such "ob- 

 servation rows." Any one of these rows could be chosen 

 for the source of such material as is required for investi- 

 gating the development of the lint, whether for pickling 

 or for testing, and reference to the routine records would 

 show at a glance the daily rate of growth in tenths of 

 millimetres, the fractional number of average daily 

 flowers or bolls, dates of irrigation, and so forth; while 



