APPENDIX 467 



permeable seeds like those of the Broad Bean [Faba vulgaris) would 

 experience a noticeable loss of solid material when swelling in water. 

 Yet Nobbe (p. iii) in his carefully guarded experiments on the 

 swelling of these particular seeds in water found that the residue left on 

 evaporating the distilled water employed amounted only to 2*3 per 

 cent of the weight of the resting seed. Immersion in water, it should 

 be remarked, is not Nature's method of procuring germination ; and no 

 doubt the loss is much exaggerated in such experiments. In my own 

 experiments, where the seed was placed in damp moss as soon as it began 

 to increase its weight in water, such a loss would be much diminished, 

 and with most hard-shelled seeds it would be almost nil. However, with 

 some seeds it is nearly impossible to obtain a good result. Thus it is 

 with those of Adenanthera pavonina^ where a gelatinous material escapes 

 from the puncture in the coats that is required to induce germination. 

 Then, one may add, it is very difficult to prevent a slight loss in the 

 case of seeds like those of Poinciana regia and of species of Guilandina^ 

 where the cuticle is apt to scale off. On the whole, however, I 

 consider that the disturbing influences in such experiments tend to 

 counterbalance each other. This remark also applies to hygroscopic 

 variation of a seed's weight in response to atmospheric changes, if 

 the seed is kept under observation for some days after the drying 

 has ceased. 



As regards the mode of conducting such experiments, it is obvious 

 that much depends on the uniformity of the conditions of drying, 

 which should be carried out in a room but Httle aft'ected by artificial 

 heat, except in the case of damp weather, when the seed may be placed 

 at first in a warm, dry room. 



Most of my results on entire seeds are included in the table 

 subjoined. The effect of hygroscopic variation is excluded by taking 

 the mean weight for a few days after the drying has reached its 

 minimum. A column is added for the periods required to reach the 

 minimum, the drying period corresponding roughly to the swelling 

 period preceding germination, since a seed that swells quickly would 

 also dry quickly. 



Conclusions to be drawn from this table are confined here to 

 immediate inferences that can be legitimately deduced from its 

 inspection. Remarks on the bearings of the data here arranged will be 

 found under the reference given at the head of the note. It is there- 

 fore sufficient to point out that although all these results indicate that 

 a seed swollen for germination returns when air-dried to approximately 

 its resting weight, there are regular variations both on the plus and 

 minus side, the ultimate weight being in excess in the case of 

 impermeable seeds, on the minus side generally with permeable seeds, 

 and variable in seeds where the permeability is a variable character. 



