48 



It must be noted that coarse grinding or the use of whole seeds hinders 

 the escape of steam evolved ; this is also the case when tising a temperature 

 of 98° C, which is under boiling point. 



For this reason a slightly higher temperature, viz., 103° C. to 105° C. 

 (electric drying oven) is being adopted in chemical laboratories in the 

 estimation of moisture in organic substances. 



A more thorough preparation of the sample, namely, grinding until 

 the whole sample passes through a sieve with a mesh of 1 mm. has 

 previously been used in these laboratories. A drawback to this more 

 thorough preparation is that it takes more time, which may allow the 

 moisture of the sample to alter during the work. 



In order to get an appreciable basis for the discussion of these questions 

 I have carried out several tests as to the influence on the results of the 

 analysis of various methods of preparation, temperatxire and time. Owing 

 to lack of time, I have been compelled to limit the analysis material, and 

 have consequently selected some of the most important kinds of seed, viz., 

 two of oats (avena sativa), two of barley (hordeum vulgare), two of rye 

 (secale cereale), two of timothy (phleum pratense), one of red clover 

 (trifolium pratense), one of alsike clover (trifolium hybridum), two of 

 turnip (brassica camp, rap.) and two of meadow foxtail (alopecurus 

 pratensis). 



These 14 samples were prepared in two different ways, namely, (1) as 

 usual in seed testing, by rough grinding the large seeds, and leaving whole 

 seeds of the small kinds, and (2) in the same way as for chemical analysis, 

 by grinding and sifting the large seeds and crushing the small seeds in a 

 mortar. These 28 samples were dried at 98° C. and at 103° C, and in both 

 cases the dryinj; was done first for 4 hours and then for an additional 

 hour. The parallel determination was carried out as far as possible, and 

 in each case 5 gr. of the substance was used. The results are shown 

 together in Table I (see page 50). In Table I, are shown the differences 

 between the parallel determinations. In Table II (see page 51) is given 

 the difference in the results for various methods of procedure. 



From Table I it will be seen that the conformity between two parallel 

 analyses is, generally speaking, good in all methods, but it is decidedly 

 better at 103° C. than it is at 98° C. ; further that the more careful pre- 

 paration also gives rather more closely corresponding results ; and that it 

 does not make much difference when one dries for the additional hour. 



From Table II it will be seen that the more careful preparation gives 

 about i per cent, more moisture than the usual method. Alopecurus is 

 an exception, giving a lower result from the careful preparation, & fact 

 that is probably due to a displacement of the moisture content during 

 the preparation. Further, drying at 103° C. gives about f per cent, higher 

 result than drying at 98 °C. Drying for the additional hour, after four 

 hours, varies the result only by a small fraction of one per cent. 



The main point is the utilisation of a method that permits of different 

 stations getting the same result. Having regard to the above experimental 

 results and to my experience as a scientific chemist and considering the 

 importance of making the method as practical and simple as possible for 

 different kinds of seed, I beg to propose that the following method be 

 adopted as international, and that it be used in all cases of international 

 transactions in seeds. 



The estimation of moisture in seeds should be carried out as follows : — 



(a) For cereals and other large seeds (dry weight of 1,000 grains more than 

 10 gr.) 5 gr. of roughly ground substance is heated for from 4 to 5 hours 

 in a drying oven to 103° C. 



Parallel determination miist be carried out. 



(6) For small seeds (dry weight of 1,000 grains less than 10 gr.) 2J gr. 

 whole seeds are heated for from 4 to 5 hours to 103° 0. Parallel determi- 

 nation as above. 



I Semarks. — -It is evident that the sample — ^when the estimation of 

 moisture is to be carried out with the object of a guarantee (and preferably 



