August 31, 1917] 



SCIENCE 



211 



suits of his own experiments in which diiler- 

 ent phosphates were compared. 



Statements- recently made by him in regard 

 to the conclusions drawn in Bulletin 90 of the 

 Tennessee Agricultural Experiment Station 

 even go so far as to impugn the ability of an 

 author who would draw the conclusion that 

 bone meal proved to be, in those experiments, 

 superior to rock phosphate. In view of the 

 detailed data contained in Bulletin 90, the 

 writer is surprised that there should be any 

 serious differences of opinion in the matter. 

 Careful consideration has convinced him that 

 Dr. Hopkins has laid unwarranted stress on 

 a single table (XIII.), which gives some av- 

 erages from the three longest-continued ex- 

 periments, and that he has failed to give due 

 weight to the results of the individual series. 

 This raises a question as to the value of such 

 a table, especially to the casual reader, for it 

 is evident that if a short number of series be 

 averaged a preponderance of a single series 

 may distort or mask the true findings. Such 

 a table, therefore, is open to criticism, and 

 evidently should be used with discretion, but 

 is justified as one way of presenting a sum- 

 mary. 



Table XIII. of Bulletin 90 gives as stated, 

 a summary from three series of experiments 

 each conducted on a different type of soil. 

 Series 1, as is pointed out on pages 69 and 70 

 and again on page 87 of the biilletin, was con- 

 ducted on a soil which proved to be naturally 

 too well supplied with phosphoric acid to be 

 at all well adapted to the comparison desired, 

 so much so that rock phosphate in the last 

 four years of the five-year period proved un- 

 profitable in three of the eight experimental 

 conditions. Excessive growth with lodging 

 reduced the yields of wheat on one half the 

 bone-meal plots, and even acid phosphate was 

 used with only a narrow margin of profit. 

 The soils of the other two series proved, how- 

 ever, to be poor in phosphoric acid and hence 

 weU suited to a comparison of phosphates. 



In series 2 the evidence is unsatisfactory 

 because of the lack of agreement between the 

 results of the two rock phosphate plots, one of 



2 ScffiNCE, November 3, 1916, p. 652. 



which shows a slight loss and the other a good 

 profit from the use of rock phosphate. If the 

 latter be compared with the near-by bone-meal 

 plots the rock phosphate shows more profit. 

 In series 3, which was conducted on a soil 

 especially poor in phosphoric acid, the ev- 

 idence is decidedly in favor of bone meal as 

 compared with rock phosphate. Under every 

 one of the four experimental conditions of this 

 series bone meal made a large increase in 

 yield — equal, in fact, to the best obtained from 

 acid phosphate and averaged 5.6 bu. of wheat 

 per acre more than that obtained from rock 

 phosphate. Even when calculated on the dol- 

 lar-investment basis used by Dr. Hopkins, the 

 average acre profit from $1.00 invested in bone 

 meal was $3.05 as compared with $2.79 for 

 rock phosphate. In this connection it should 

 be mentioned that a comparison between bone 

 meal and rock phosphate where the cowpeas 

 were removed for hay was omitted in Table 

 XIII. because only in series 3 was such a 

 comparison made, the results being especially 

 favorable to bone meal. 



Series 4, which was not included in Table 

 XIII., is also worthy of consideration. This 

 series was conducted on a greatly impover- 

 ished type of soil, well known to be naturally 

 poor in phosphoric acid. As measured by the 

 yields of wheat, acid phosphate proved highly 

 profitable, but both bone meal and rock phos- 

 phate were used at a loss. However, the 

 writer's records and observations of these ex- 

 periments, during the two years of their con- 

 tinuance, convinced him that bone meal could 

 be used profitably in the reclamation of land 

 of this character. On the other hand, rock 

 phosphate appeared next to worthless. By 

 way of confirmation, bone meal plots 9 and 11 

 produced in the second year an average of 

 1.41 ton of cowpea hay to an acre. The near- 

 by rock phosphate plots 7 and 8 produced only 

 0.80 ton. The value of the difference between 

 the two yields of hay would pay for the bone 

 meal used and leave a good profit. The hay 

 data were not given in Bulletin 90, but serve 

 as a good illustration of the advantage in the 

 interpretation of results that rests with the 

 person conducting the experiments. 



