Seed Wheat. 33 



the length of the rows was reduced at harvest time by cutting out the doubtful parts. 

 There were but few cases where this discretion had to be exercised ; it was exercised as 

 sparingly as possible. The case had to be one of the most manifest injustice to one or 

 the other of the contestants before I ordered any interference. It Will be seen from the 

 measurements of the rows (the normal length was about 192 '5 links), this discretion was 

 not often exercised. It should be added that in most of the cases where the row harvested 

 is less than 192 '5 links long, it is because there was insufficient seed of one of the con- 

 testants to fill a whole row. This will be better understood by reference to the diagram 

 illustrative of the arrangement of the experiments. (Fig. 28, p. 24.) The fragments of 

 rows were due to the fact that the supply of seed did not fill exactly a number of full 

 rows, and the seed was too valuable to be wasted, its production having cost a considerable 

 amount of expert labour. 



It will be seen that the results given in the tabulated weighings of grain and straw 

 relate solely to the plants that grew, this being probably the test that most nearly meets 

 the requirements of practice. I have recorded the number of plants that grew in each 

 case, but have concluded, for the present at any rate, to make no attempt to interpret 

 the figures. 



The conditions of the trials were field conditions, modified to meet the requirements 

 of exact experiment. That the yields exceed those of actual agricultural practice of the 

 region where they were conducted is due to careful planting in measured drills and to 

 the attention already described as having been given to the prevention of weeds. 



Returning now to the percentages of growth, it will be seen that from the 325 grade 

 down there is a regular diminution in the power of the seed to produce plants under field 

 conditions. The number of plants is so great, running as it does into thousands, and the 

 number of varieties is so great, reaching as it does above twenty widely different sorts, and 

 the dates of sowing are so various and the soil conditions are so varied, including as they 

 do patchiness and a general variation from one end of the area to the other in the manner 

 described elsewhere, that this variation in the power to produce plants appears to be 

 proved to be a function of the size of the seed. So far as one season's work can go, I do 

 not see how it is possible to reach any other conclusion than that the inherent power in a 

 grain of wheat to produce a yield of grain or straw is some direct function of the size of 

 the grain, and that the larger the grain the better the resulting plant will be, independent 

 of any conditions that appeared during the particular season under discussion. The 

 number of trials of the 325 and 225 grades is rather too few to be quite satisfactory ; and 

 I am inclined to think that a larger number of trials of these grades might have slightly 

 modified the percentages of growth of these grades as given. 



At this point, I would like to once more call attention to the fact that the conditions 

 of the experiment were such as make me believe that the small, and therefore weak seeds, 

 and the shrivelled and therefore weak seeds, were decidedly favoured. The careful 

 planting and the keeping down of the weeds would, it seems to me, give relatively greater 

 benefit to the weaklings than to the others. It is, therefore, reasonable to suppose that 

 under ordinary field conditions, as, for instance, when the grain is broadcast and the 

 weeds are given full swing, the percentages of growth would be lower still for the 

 smaller grades than here recorded, lower in reality and lower in proportion. 



I must remark, for the benefit of those who have never had an opportunity to observe 

 the behaviour of wheat plants growing in drills, that the interpretation of the numbers 

 given in the tables needs to make allowance for a certain extra growth due to lack of 

 competition when a plant fails anywhere in the drill. In such case, the two plants next 

 the vacant space are given more soil room and are freed from the efforts of a competitor 

 on one side, and they consequently often grow a little better. This is an observation easily 

 made, and the fact is well known no doubt to many, but it might be overlooked by others 

 if no mention was made of it. This factor may have helped the rows with the greater 

 number of " misses " more than it helped the other rows. It certainly would do so in 

 some instances. 



Thus, when it is said that one grade yielded 178 plants and the adjacent grade yielded 



nly 163 plants, it must not be assumed that the yields would tend to become as 178 to 



163, for this is so only to a certain extent, owing to the fact that the plants in the 



mailer lot profited considerably more by the death of their companions in the same row. 



Doubtless, also, this same influence extended from one row to another, but it was only 

 to a very limited extent. It was, in fact, so small and so uncertain a factor that I had 

 to abandon all attempts to settle its value. It was most certainly very small too small 

 to exert an influence that would be more than barely appreciable. It is, in fact, quite 

 remarkable how little influence even powerful factors exert beyond the distance of a few 

 inches in most seasons in the soil used for these experiments. For instance, a heavily- 

 manured row will fail to have more than a very faint effect on a check row two links 

 away and without manure. 



