072 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



analysis of some of the borings. This is rarely done, hence hundreds of 

 field-plot experiments are useless. The author advises rejection of all 

 experiments in which soil investigations have not been made, and in 

 which parallel plots are not used. 



(3) Statistical Method. — By this is meant the averaging of as many 

 field-plot experiments as possible, carried out on one plan, in as many 

 localities as possible. The method includes three distinct phases : (a) A 

 specialist prepares an experimental plan and instructions ; (b) the farmers 

 in different localities provide the ground, apply manures, and grow the 

 crop ; (c) the specialist gathers the crop-returns together, and reckons the 

 average value of each manure used. The method can only be successful 

 if each part is thoroughly carried out. The farmer's part is very liable to 

 error, for many reasons, but one would expect that the specialist would be 

 reliable. Dr. Baumann, however, shows that even here there is much to 

 be desired. To show the line of criticism we take one case — a plan 

 prepared in the Agricultural Polytechnic at Berlin. The experiments are 

 intended to compare the value of three phosphate manures : super- 

 phosphate (phosphate and sulphate of calcium), Thomas phosphate 

 (phosphate and hydrate of calcium), and Algerian phosphate (phosphate 

 and carbonate of calcium). It will be seen that the accompanying salt of 

 calcium varies in each case, yet the plan of experiments takes no account 

 of this. From the results obtained it is impossible to see whether the 

 phosphoric acid or the calcium salt was the chief agent. The plan was 

 taken up by several farmers, but in no case was an examination of the 

 soil made. The result was that with each of these phosphates the harvest 

 was better in half the cases than the unmanured plot, but in about half 

 the yield was worse. And here comes in the "statistical method." By 

 estimating the average of the harvest results in different ways, one may 

 demonstrate one of three consequences : (a) phosphate manures repay 

 many times over their original cost ; (b) they just repay their cost ; (c) they 

 give worse results than if they had not been used. Statistics are here 

 (the author remarks) the greatest liars in existence. Evidently much 

 remains before this method is perfected, not only in Germany, but also in 

 Britain. The author proposes to show how improvement can be effected 

 in a future series of papers. — W. G. S. 



Marshall Islands, The Flora of the. By G. Volkens (Not. Kdnig. 

 Bot. Berlin, Bd. IV., pp. 83-91 ; August 30, 1903). — A list of Schnee's, 

 Finsch's, and Jensen's plants of this relatively poor flora. — H. M. W. 



Masdevallia Laucheana. By R. A. Rolfe (Orch. Bev. Nov. 1903, 

 p. 847). — Historical and descriptive particulars of this pretty small- 

 flowering Masdevallia are given. — H. J. C. 



Maxillaria funerea. By R. A. R. (Orch. Bev. Aug. 1903, 

 p. 232). — Descriptive particulars of this species are here given from a 

 plant that had recently flowered in the Royal Botanic Gardens, 

 Glasnevin. — II. J. C. 



Maxillaria gracilis. By R. A. R. (Orch. Bev. Aug. 1903, 

 p. 230). — Descriptive particulars of this Brazilian species. — II. J. C. 



