860 The Sampling of Manures and Feeding-sht^s. 
turned over and mixed thoroughly, divided as before, and so on as 
may be necessary, until a quantity weighing only about three or 
four pounds is left. Two well-fitting tins, each capable of holding 
from ^ lb. to 1 lb. of the material, should then be filled from the 
heap thus left. One of these should be wrapped up and sent by 
post to the analyst, and the other be kept by the farmer for 
reference. Instead of a tin, a wide-mouthed glass bottle with 
well-fitting cork may be used, and this be enclosed in a wooden box, 
and so be sent by post or rail ; or, the sample may be wrapped in 
tinfoil or in oiled silk, and be enclosed in a box or in a stout linen- 
lined envelope. This latter is a very convenient form for nitrate of 
soda, sulphate of ammonia, kainit, and similar salts. The tin is, on 
the whole, the most satisfactory, as it is easy to send h lb. or 1 11). of 
the manure in it, whilst if a bottle, or tinfoil, or oiled silk be used it 
is not easy to send so large a quantity. If a smaller quantity be sent, 
the final heap must be mixed still more carefully, and the sample be 
taken from different portions of it. In no case, however, should less 
than 4 oz. be sent as a sample, and when the material is at all 
uneven in character, or lumpy, or of a mixed nature, it is not satis- 
factory unless a 1 lb. sample, or in some cases as much as 2 lb. be sent. 
The more uneven the manure the larger the sample must be ; the 
finer and more even it is, the smaller may be the quantity to be sent 
for analysis. 
One caution further is necessary. Whilst care must be taken to 
ensure a fair sample being drawn, care must also be exercised not to 
let the portion that is being sampled lie about exposed too long. 
The sampling must be done carefully but also quickly, or the material 
may dry considerably during the process. 
In the absence of a special sampling tool such as that described, 
the best way is, after selecting several bags as directed, either to 
turn them out one after the other on a fioor, and, taking a few- 
shovelfuls from each, to mix these shovelfuls well together for one 
sample, or (which is not so good) to drive a spade into each of the 
selected bags, and, after a little mixing, to draw out from as near the 
centre as possible a couple of spadefuls from each bag, subsequently 
mixing these lots together, flattening the lumps down, and dividing 
and subdividing the heap until only some three or four pounds are 
left. From this the tins or bottles may be filled as mentioned before, 
one sample being sent for analysis and the other retained for 
reference. 
There remains the special case of the sampling of bulky materials 
such as shoddy, wool refuse, &c., wliich, as a rule, are sold, not in 
bags but in bulk. In such cases a " sampler " is of little use, and 
the only way is to take portions from different parts of the bulk, 
or from the selected bags (if bags there be), and, setting these 
portions aside, to quickly tear apart the matted wool, hair, ttc, 
and mix the lots together as well as it is possible to do, but losing 
no time so that the material may not dry unduly in the mixing. 
The final sample must be taken from this by rapid subdivision 
and remixing as before, but a considerably larger quantity must 
