Page 334 Oil yields varied in the same way as kernel yields, the average at the first planting being 544 lbs. per acre and at the last 220 lbs. per acre.
At Urambo, experiments on both red and grey soil compared mid-season and late plantings of Virginia Bunch, Natal Common, Spanish Bunch and Valencia groundnuts. The relevant dates were 6th and 16th February for early and late plantings on the grey soil, and 14th January and 11th February for early and late plantings on the red soil. All these plantings except the earlier one on the red soil were later than any at Kongwa, and too late for satisfactory yields, particularly in a season in which February was distinctly dry. Even so, the decline with later planting was shown for all varieties. At the early date, yields varied on the red soil from 600 lbs. kernels per acre (Spanish) to 1,000 lbs. (Valencia), while at the later date the best yield was under 400 lbs. (Valencia). On the grey soil, Valencia, the best yielder, pro- duced only 400 lbs. at the early planting, and 300 lbs. at the later.
When the effects of varying plant populations are roughly eliminated from the yields, by the somewhat crude device of calculating the performance values (yield per 1,000 plants), and these values are plotted against planting date, characteristic falling curves result for all varieties (Fig. II). The mean curve for Urambo (neglecting for this purpose the difference between red and grey soil) has the same slope as the mean Kongwa
Kongwa curve, indicating that the planting date/yield relationship is to some extent independent of the climatic difference between Urambo and Kongwa (section 12). It is therefore reason- able to suppose that earlier plantings at Urambo would have given consider- ably better yields, particularly since good rains fell there in late December and January. This also explains the poorer results generally obtained on the grey soil on which no plantings were possible owing to excess of water until 29th January.
Cultivation methods and harvesting
The effect of ridging was tested at Kongwa in a hand-planted experiment which compared ridge planting, flat planting and planting in the furrow. (Other treatments, whose effects were negligible in this experiment, tested liming at different rates and broadcast application of P2O..) The results were as follows in lbs. kernels per acre adjusted for plant number variations:
Ridge planting Flat planting Furrow planting
•
·
Standard Error ± 34.8 lbs.
1,161 1,028
786
Ridge planting is significantly better than flat planting and flat planting is significantly better than furrow planting. The ridged plants escaped the water- logging in early stages of growth which affected the flat and furrow planted plots. It would, however, be unwise to recommend ridge planting on the results of a single experiment, particularly as it would make impossible the very valuable early harrowings on the flat.
Late ridging on Virginia Bunch nuts, at Kongwa, had no effect on yields but facilitated harvesting.
In addition to improving yields, early ridging after planting and harrow- ing on the ffat will probably be of value in aiding harvesting on compacting soils, such as the Kongwa red type, and will help in water conservation and in erosion control, particularly when it is combined with tie ridging (basin listing), a practice which under dry land conditions at Ukiriguru in the Lake Province of Tanganyika, has given striking benefits. Further, in the wetter areas ridging may well help to reduce the dangers of excess water. So far as can be seen at present, the method of planting outlined above, using the twin row technique on the flat, should be immediately preceded by a light harrowing and immediately followed by another. After ten days or so, a further light harrow- ing should be used to knock down weeds, and subsequent cultivation should ridge up the double rows, so that they come to harvest on either side of a ridge. Harvesting would then consist of breaking down the ridge rather than as at
Page 334
114
Page 334
176
1
*
tryb Page 335 of 1097