SELECTION OF SEED. 221 



measure which did not "weigh as much as at least 270 

 grains/' 



Of course, in estimating the weight of samples, says the 

 author (Professor Buckman), it is always necessary to see 

 that it is made up of the proper seed, as, if it be full of 

 plump grains ofBromus mollis (soft brome or lop), Triti- 

 cum repens (couch grass), or, as is frequently done purposely 

 to add to the weight, of Plantago lanceolate (rib grass), 

 its weight will be greatly increased, and unfortunately 

 in such a way that the grass will be too poor to grow, 

 but the weeds will be sure to succeed. We have seen 

 fields sown with ryegrass, in which the only crop visible 

 has been plantain or rib grass; and we must recollect that 

 both plantain and brome grass are strictly annual plants, 

 and will have sown their seeds freely, while even the 

 barley with which the "seeds" may have been sown 

 down is being harvested. The point we would here 

 advert to is, that the table demonstrates that for the 

 most part the lighter samples of seed are fullest of weeds. 

 This may arise from the carelessness of its growth on the 

 one hand, and the impossibility of cleaning mere chaffy 

 stuff on the other. As may be expected, the heavier 

 samples of seed have been more carefully grown, it being 

 next to impossible to obtain the best samples by a dirty 

 cultivation; and then, again, the heavy grains are more 

 plump and rounded, the chaff fits to the seed, is not so 

 pointed, and therefore not so liable to entanglement, so 

 that these can be separated with tolerable ease. That 

 clean samples can be obtained is obvious from our samples 

 10, 12, and 14, in which the weeds were remarkably few 

 in number, and such as could readily be separated. That 

 the heavier samples of ryegrass seed are the cheaper we 

 can state from repeated experiments, as we have sown 

 the lighter ones, in which scarcely a genuine seed would 

 germinate, whilst of the heavier samples only a slight 



