862 Tlie Sampling of Manures and Feeding-stuffs. 
the first, that the sample sent be fairly representative of the bulk ; 
the second, that it suffer no change in transit. 
Feeding-stuffs. — With these, as with manures, similar precautions 
must be taken for the obtaining of a representatiA'e sample, but the 
changes that are likely to happen in transit are of but small conse- 
quence, as a rule. In their place, however, comes in a necessary 
caution as to the particular obserA"ance of the condition of a 
delivery. 
Those who have had experience in analysing feeding cakes know 
that, even in a single cake, results differing to some extent one from 
the other may be found, according as the piece which is taken for 
analysis may happen to have been broken off the end, the middle, 
or some intermediate part of a cake ; foi-, owing to the way in which 
the crushed seed is packed in the oil-presses, it is impossible to get 
the pressure equally distributed throughout. Some parts of the 
cake, as evei'yone is aware, are generally thicker than others. The 
oil is pressed out from the centre and runs away fi-om the edges of 
the cake ; finally, the edges are pared off all round. If a single cake 
be taken, the best way is, either to send up for analysis half the 
cake or to break a whole cake across the middle, and then off each of 
the halves to take a strip about four inches wide also right across 
the cake and from what was before the middle piece of the whole 
cake. Thus two good-sized pieces, taken across the whole breadth 
of the cake, will be obtained, and they can be wrapped in paper and 
sent by post or rail. But even this is not fully satisfactory, for 
sometimes a delivery of cake may vary ; more especially is this the 
case with decorticated cotton-cake. Occasionally, too, some cakes of 
a delivery may be a little mouldy or damaged, but a single cake of 
this kind may not be faii'ly representative of the whole. It is quite 
true that if a delivery of cake be guaranteed to be pure, every cake 
of the whole lot ought to conform to that description, but when the 
oil percentage or some other constituent, or the general condition 
be in question, a fair sample of the whole should be taken. A 
purchaser should, therefore, first look over the cakes comprising the 
delivery and note any differences of appearance that may strike him, 
or see whether all the cakes seem much alike. He should then select 
samples from each different variety he notices, the number of samples 
being in proportion to the number of cakes of each kind that make 
up the bulk. Thus, if about half the delivery consists of darker- 
coloured cake and half of lighter-coloured cake, an equal number of 
cakes should be taken as samples to represent each sort. Three or 
four cakes of each sort should be selected, or, if uniform throughout, 
say, six cakes from the whole lot ; pieces should be broken out of 
the middle of each as described before, and then these pieces be 
passed through a cake-breaker. The broken " nuts " or lumps must 
next be mixed up thoroughly, and then divided successively, just 
as was advised in the case of manures, until only about a couple of 
pounds weight are left. Two tins may now be filled with the cake,- 
one for sending to the analyst, the other to be kept for reference. 
Or, the pieces may be done up in a bag, or wrapped in stout 
