A METHOD OF YELD ESTIMATION. 



By A. 0. D. MoGG, B.A.. 

 EcuJogist, Division of Veterinary ResearcJi. 



(Abstract.) 



BeaJ Jul J, IT, 1920. 



This method originated from the necessity of making a 

 very close study of certain stock disease-bearing velds occur- 

 ring mainly in Natal, Zululand and East Griqiialand, the 

 diseases being caused, in the opinion of various veterinary 

 collaborators, by some plant, plants, or even a condition of the 

 veld herbage. 



Several diseases were investigated, of which the principal 

 are .stijfziekte, dunziekte and cattle pushing-staggers, the 

 latter now happily solved by the application of the method. 



The problem set to the botanist was to find the plant, and, 

 having adduced enough evidence to incriminate any plant, it 

 was then necessary to prove this to be the plant cause by apply- 

 ing the crucial test of feeding the plant to animals and 

 producing in them the typical disease. This latter test, of 

 course, was carried out by the veterinarian. 



Initially, the method adopted was to list separately all the 

 species occurring on as many intensely "diseased " areas as 

 possible, choosing those farms most widely separated, in order 

 to secure geological and climatic diiferences as well. 



The plant-lists were arranged both ecologically and 

 systematically for each farm. These local lists were then 

 compared for all areas carrying* the same disease. At once 

 the eliminative effect of such a process can be seen to give one 

 relatively small common-factor lists, the size of whi(-h being 

 of practical dimensions, enabled one to test ea(h plant 

 separately. However, a further use was made of these lists. 

 As the lists recorded the range of variation of a species in 

 form, habitat and distribution on intensely " diseased " 

 areas, it was sought to ascertain whether there was in these 

 data any correlation to the disease, i.e., any ratio or other 

 factor, such as the relation of genera to species, common 

 genera to the common species, etc. 



To test this, it was easy to count the species, genera and 

 orders for each locality, reducing all ratios to percentages and 

 comparing them in pairs. Some interesting results were 

 obtained, and I give the final extract figures for four widely 

 separated dunziekte areas in illustration. 



(a) Total Genera : Total Species: — 'Mooi River and 

 Kokstad, 444: 55-6; Rosetta and Kokstad, 417: 58?,: Molteno 



