R. N. 8ALAMAN 1.S9 



overlooked that in a normal season the sets weighing 2 ozs. or thereabouts 

 might again prove to be the best, and that the advantage accruing to 

 heavier sets in 1921 is due indirectly to the great drought, when the 

 absence of moisture in the soil gave a fictitious value to tlie heavier tuber- 

 sets because of the greater quantity of moisture contained in and about 

 their relatively larger bulk. 



The lots H, J and K. representing the tuber-sets classes with out- 

 growths, produced crops of 50, (il-6 and 50 lbs. respectively. In the 1920 

 trials the outgrowth tuber-sets were whole and weighed 6 ozs. each, and 

 produced a crop 25 per cent, bigger than any other class of seed. In this 

 year's experiment there were no large tubers with outgrowths. Indeed, 

 the latter, on any size tubers were rare. 



The outstanding feature is the high yield of class J, the cut sets with 

 outgrowths of average weight 2 ozs. This exceeds the yield of H or K 

 by 14-6, which is five times the probable error for a single pair, and 

 although it is true that the probable error for a plot of 65 tuber-sets is 

 greater than for one of 100, yet a difference as great as here observed 

 cannot be regarded as other than very significant. 



Assuming, therefore, as we may. that the excess of yield of the .T 

 class over that of H or K is to be referred to the difference of kind in 

 the seed set, it becomes of interest to consider what this difference really 

 consists in. 



In class .1, the only buds left are those eyes on the secondary out- 

 growth, whilst there are 2 ozs. of flesh for their shoots to feed on in their 

 early stage. In class H we have whole tubers of the same weight as the 

 J class, but here the outgrowth is only a part, and that a minor one, of 

 the whole tuber, and the shoots arising from it have to compete for 

 nourishment with those arising from the other eyes of the tuber, and 

 hence have not the initial advantage which the J class enjoys. 



Class K consists of isolated outgrowths with but small quantities of 

 tuber material to feed on and is, in a season such as that of 1921, at a 

 distinct disadvantage. 



Class I consists of the rose or crown ends of the tubers from which 

 the cut sets of J and K have been removed. Their average weight was 

 1-66 ozs. There is no reason to expect that these sets should possess any 

 greater advantage than any other tuber-set of the same size which was 

 devoid of outgrowth. Indeed, in respect to weight of tuber-set the yield 

 that might be expected would be something between that given by classes 

 B and C in the same row, say 52 lbs. However, the reahsed yield of 40 lbs. 

 is sufiicientlv widelv removed from this figure to suggest a difference of 



