150 TESTING OF SEEDS. 



HAOTTEMAKNT'S GERMINATION PLATTER [Figure 6). This is a cir- 

 cular plate of porous earthenware about six inches in diameter 

 and two inches thick. On the upper side are regularly arranged 

 circular hollows numbered consecutively from 1 upwards, and about 

 in. across and o/lG in. deep. The underside bears eight 

 radial furrows about 1/5 in. wide and 3/25 in. deep, which facili- 

 tate the passage of water through the earthenware, from the under 

 to the upper surface. In using the platter, it is first soaked in 

 water for about 24 hours and then placed in a flat dish containing 

 just enough water to reach its upper edge without flooding it. The 

 seeds are placed singly in the hollows, and either left uncovered or 

 covered over with flannel or good soil. The water should be re- 

 plenished whenever necessary. 



NOBBE'S APPARATUS (Figure 7). This consists of a square- 

 s_haped block of porous earthenware of about eight inches side and 

 two inches thick and with the under surface and sides glazed. On 

 the upper side is a saucer-like hollow surrounded by a canal, the 

 depth of which is almost equal to the thickness of the block. The 

 seeds are placed in the bowl and are kept constantly moist by the 

 water which percolates in from the surrounding canal through the 

 porous earthenware. To keep out external influences, the appara- 

 tus is protected with a cover (c in figure), which is raised at the 

 corners to allow a free circulation of air, and is perforated in the 

 centre to admit a thermometer in case of delicate experiments. 



Experiments carried out by Baur prove that all the various 

 methods of test by germination yield more or less the same results, 

 although, on the whole, sowing in a garden bed is the least to be 

 relied on, while Nobbe's apparatus is the safest. 



The germination test is the best of all, as it is the most certain 

 of the three, and by it alone can the exact proportion of good seed 

 and the germinative power of the seeds be accurately gauged. But 

 it is evident that it can be employed only with seeds that can 

 be forced to germinate within a reasonable time ; otherwise 

 the test may be vitiated by unavoidable accidents during the long 

 interval between the commencement and the close of an experi- 

 ment. Moreover, there are some seeds, like those of teak and 

 babul in many parts of India, an entire lot of which, all collected 

 at one and the same time and from one and the same tree, never 

 germinate together but at intervals of weeks, mouths and even one 

 and two years. For sjuch seeds the germination test alone would 

 necessarily be insufficient and in most cases impracticable, but, 



