204 EXPERIMENTAL STATIONS. 



acre is the size which has been adopted as possessing the greatest 

 advantages. There are few advantages gained by the use of 

 small plots that cannot practically be attained on rood-plots, and 

 there are some special reasons why the latter are to be preferred. 

 Chief among these is the possibility of conducting the experi- 

 ments upon an agricultural scale without interfering with the 

 ordinary methods of farming. Whether this makes any very 

 essential difference to the success of experiments which are not 

 of a mechanical, manipulative kind, is difficult to say. Spade 

 culture is said to be more favourable to tlie growth of some 

 crops, such as clover and leguminous plants, than plough culti- 

 vation, and this may probably be due to the deeper disturb- 

 ance and more thorough mixing of the soil; but whether 

 this is so or not, it is certain that farmers are somewhat 

 sceptical of results which have not been obtained by ordinary 

 methods of cultivation. The plots on the Society's stations 

 are all nearly 400 feet long and 28-|- feet broad, and, as will be 

 seen from the accompanying diagrams, are so arranged as to give 

 a clear run for the plough of about 800 feet. The size of an 

 experimental station that is all under one crop should not be so 

 large as to prevent its being sown in one day, for a difference of 

 a day or two in the time of sowing has a marked effiect on the 

 crop. This has been very clearly brought out at both stations 

 during the past season, where, from circumstances which will be 

 avoided hereafter, the sowing at Harelaw was not completed in 

 one day, and the occurrence of wet weather prevented its being 

 resumed for nearly a fortnight. The consequence is that the 

 plots first sown took the lead of the other plots, and kept it to 

 the end. An objection has been made to large plots that their size 

 is such as to prevent the employment of a sufficient number of 

 duplicates. This objection vanishes when we consider that each 

 plot is capable of being divided up into twenty-eight plots — each 

 the rla^^^ ^^ ^^ ^^"^^ ' ^^ ^^^^^ e?ic\i rood-plot really yields a crop 

 which would have been reckoned from the average of twenty- 

 eight such small plots, though the probability is that any one of 

 these would not have produced a crop exactly the twenty-eighth 

 part of the rood-crop. Even in a well-selected field, the varieties 

 of soil which may occur in the space of one rood is surprising. 

 The station at Pumpherston affords striking evidence of this. 

 During the autumn there was scarcely a plot where a practical 

 eye could not detect differences in the appearance of the turnip 

 crop which seemed to be due to that cause. Iq some cases this 

 has occurred to such an extent as to render it questionable 

 whether certain plots are capable of being used for comparative 

 experiments. To enable an estimate to be made of the various 

 grades of fertility of the different parts of a field intended for 

 experiment, it would be advisable to sow it for one or two 



