OECOLOGY 0 E SOME PLANT COMMUNITIES 35
to be taking place in tbe distribution of tbe species. It has always, during these years, shown a fairly sharp distinction into four parts passing from North-west to South-east and approaching the bank of the Sabarmati River. In some cases the dividing line between the communities is very abrupt but between the third and fourth parts it is more gradual.
It was apparent from the small size of the whole area, which is about half a mile from North-west to South-east, and about 100 yards from North-east to South -west, that no perceptible differences exist in the climatic factors, humidity of the air, light and tempera- ture. The altitude was also practically constant except for a very slight depression in the second part. It seemed therefore that the area was admirably adapted to an investigation of the differences between these four parts and their causes.
A preliminary general account of the oecology of Northern Gujarat was published by Saxton and Sedgwick (7) in 1918 in which a description of the Savannah associations is given. In that paper, the area we have studied is included under the heading of “other types of Savannah on Sandy Soil " (loc. cit. p. 233) and its relation to other types is shown.
2. Investigation
A . General . — The area selected is a plot of waste ground about 3 miles South-west of Ahmedabad on the North-west bank of the Sabarmati River. Its approximate dimensions have already been mentioned. Eor the greater part of its length it is bounded both North-east and South-west by cultivated land, but to the South-east it merges in a stripe of waste land extending along the river bank.
The four parts of the area, mentioned above, will be distinguished as Areas (I), (II), (III) and (IV), passing from North-west (l) fc to South-east (IV).
Area I is largely dominated by Cassia auriculata Zb, Area II by Acacia arabica Willd . and Cassia auriculata Z/. in similar proportions, Areas III and IV by Saccharum Munja Roxb., its dominance being more marked in Area IV.
In order to arrive at any definite conclusion five lines of research seemed to be desirable (a) a detailed analysis of the vegetation of the areas ; (b) an analysis, chemical and physical, of the soils, including the water contents; (c) a study of the physiological anatomy of the plants concerned ; (d) an investigation of the root systems of some of the more dominant plants and (e) experimental determination of the wilting point of the important plants, and their rate of transpiration
