THE CONTROL OF THE FACTORY 



549 



They must therefore be sampled in numerous places, samples being taken 

 also from many cakes. 



Sampling of Bagasse. — The sampling of bagasse, which is very important, 

 is also the most unsatisfactor}^ problem met with. It is of unequal composi- 

 tion due to the structure of the cane, to unequal distribution of added water, 

 and to inferior crushing at the extreme ends of the rollers. To avoid error 

 from these causes the sample should be taken from across the whole width 

 of the rollers. The subsequent treatment depends on the method of analysis 

 used. If small quantities — lOO grams — are used in the analysis, it is im- 

 perative that a large sample of, say, a kilogram be chopped to a fine meal 

 in some machine, such as a sausage-meat chopper. This process is trouble- 

 some and invariably entails some alteration in the composition of the matericil. 

 It is much better to make the analyses on a larger quantity, say, one kilogram, 

 and to avoid the sub-sampling. With efficient modern mUling, bagasse is 

 in a suitable condition for analysis without further division. 



Bagasse taken from the earlier mills of a train for special analyses must, 

 of course, be reduced to a fine state of division. 



Fig. 354 



The sampling of bagasse cannot be automatic, nor 3'et can it be safety 

 preser\'ed for anal5'sis. Its composition depends on the feed of cane and on 

 the quantity of water used. The samples should then be taken under normal 

 working conditions and should indicate as the result of their analysis what 

 has been the average, and not what was the composition of bagasse at any 

 particular moment. In the system of operating cane sugar houses lack of 

 appreciation of this point often leads to friction between the engineer and 

 the chemist, both often forgetting that they are merely individual units in 

 a complicated machine. 



The nmuber of samples and analyses necessary to obtain an average 

 result reasonably accurate will depend on the variation between individual 

 analyses, and this variation will depend on the variation in the raw material, 

 the regularit}' of feed, and the general oversight exercised on the operation 

 of milling. An hourly or at the least a two-hourly sample and analj'sis is 

 generally necessary. 



Sampling of Cane. — In general cane cannot be satisfactorily' sampled 

 since the variation from stalk to stalk is great, and also the composition of 

 individual stalks varies from butt to top. \Vhen circumstances arise such 



