VELD ESTIMATION. 225 



Examination of these figures will at once slnnv : — 



1. The enormous drop in the percentage proportion of the 

 best grazing grass, AntJii.^flria imberhis, the " blue grass " 

 of Natal, by the veld treatment accorded. 



2. The increase in proportion of useless grasses, notably 

 Elionui'us orgenfeus, to replace Anthistiria. {EHonurns was 

 not only avoided by the " army worm " when all else was 

 eaten, but cattle tethered on this grass rather starved tlu^i! 

 ate it.) 



By making graphs there is apparently an intimate relation 

 between Elionurus (with one or two other grasses) 'nd 

 Antliistiria. A distinct correlation, as set forth in (3), seems 

 indicated. 



3. That an exact estimate of the grazing value of each 

 paddock can be calculated. Paddock A is a typical sample 

 o'f the good grazing in central Natal. Paddock B, adjoining, 

 is now a poverty veld; B's cattle grazing* this paddock never 

 thrived on it, and contracted stijfziekte. Our experimental 

 cattle have contracted stijfziekte in as short a time as three 

 weeks after being removed to Paddock B, ha^ang grazed in 

 Paddock C for over a year./ The method thus gives us a 

 (quantitative estimate of details of this description. The results 

 are valuable, as there is no reason to suppose that thirty years 

 ago the three paddocks were not identically similar in average 

 botanical composition. 



But the method has a fairly general application in the 

 exact detailed estimation of (1) veld change, i9\ veld compo- 

 sition, (3) distribution of vegetation — that is, ecological values. 



I will briefly indicate some of the uses : ■ — 



1. The quantitative estimation of the grazing value of any 

 farm at a particular time for land valuation, and, by continu- 

 ous observations, the effect of such factors as season, burning, 

 sto(d\ing, drought, ploughing, etc., could be estimated. 



2. The quantitative and qualitative estimation of the rela- 

 tion of vegetation to the soil by comparing smdi belt transect 

 samples taken from different localities bearing the same soil. 

 As such estimations can be carried out with extreme accuracy, 

 such a method used in conjunction with soil survey and 

 analysis would yield invaluable re(tords. 



3. Estimation of the actual, and i)rediction of the 

 possible, effect of burning of any specific veld, especially if 

 continuous burning is practised. All my graphs of burned 

 veld show this remarkably clearly. 



4. In a transect taken from ridge downwards through a 

 marsh most striking results were obtained. These go to show 

 that (a) plants exhibit a very definite zonation iu their 

 euhabitat. Thus, if a tacheometrical survey of such an area 

 were made, taking vertical intervals of 1 foot on a slope of 

 about an angle of 10 degrees, then each contour line plotted 

 would represent almost exactly- the zonation exhibited by a 

 different species in the descent from ridge to vlei. (b) A com- 



