The Kitale Maize Agronomy Programme, 1963-75.
Dr Alister Allan

Early research work in the Kenya Agric. Dept. concentrated on cash crops like coffee, wheat, tea and pyrethrum, while maize experiments were carried out by researchers on district stations, which covered several crops as well as livestock. Pre-1963, there had been 64 trials on fertilizers, 34 on varieties, 14 on spacing, 18 on cultural and weeding practices, and 2 on time of planting. These trials, along with the practical experience of growers, had built up enough information for making general recommendations. However, maize was a relatively new crop in the Highlands, both to African subsistence growers and to the European commercial farmers. Previous trials were mainly mono-factorial, hence gave no information about interactions, and there were no data on the relative importances of factors affecting yields under different conditions. Finally, judging by experience in other countries which had introduced hybrids, the imminent release of  Kenya's first hybrids in 1964 was likely to shake up the whole maize-growing system. Therefore, at an early stage in the Kitale agronomy programme, a series of multifactorial District Husbandry Trials DHT was started to get reliable data on the relative importances of six factors thought to be the most important ones affecting farmers yields; obtaining data on interactions was also essential. Each factor was at either a High level, representing current good practice, or at a defined Low level of poor practice. The treatments were:-

The Design of the Experiment

FACTOR	HIGH LEVEL	LOW LEVEL
T -Time of Planting	Early in rains	4 weeks late
S - Plants per Ha	40,000	20,000
V - Type of seed	Hybrid	Local
W - Weed Control	Early & clean	Late
P- P2O5 Kg/Ha	56	0
N - Nitrogen Kg/Ha	80	0

The design was a 2x2x2x2x2x2 factorial with 64 combinations and 1 replication/site. These DHTs were laid down at several sites around Kenya;results everywhere were similar, but only those from 3 sites around Kitale in 1966, repeated in 1967, are reported here. Yields, in Q. (100Kg)per Ha, from the main treatments were:-

The Experimental Results

Factor	Low	High	Diffcrence	Statistical Significance
T	41.2	61.8	20.6	***
S	44.9	58.1	13.2	***
V	42.9	60.1	17.2	***
W	46.2	56.8	10.6	***
P	50.6	52.4	1.8	-
N	48.3	54.7	6.4	***

T and V had relatively the largest effects, followed by S and W, while fertilisers had less effect overall. Clearly, farmers would increase yields most by concentrating firstly on good husbandry, i.e. High T, S and W. Although Hybrid increased yields under all conditions, it gave bigger increases with good husbandry. There were many significant and important interactions, particularly TN, TW, TP, SV, SN and PN, as well as TWN and WPN. The low overall fertilizer response was due to interactions between fertilizers and husbandry - with poor husbandry, fertilizers have little effect, whereas with good practices, they increase yields profitably.All the significant effects can be incorporated in an equation which calculates the yield for any particular treatment, e.g. All High = 90.5 and All Low =21.1.The data can be used to work out the profitabilities of different treatments, to compare the relative efficiencies of different combinations of the six factors for improving yields, e.g. the effects of using Good Husbandry (High TSW) with local seed and no fertilizers (yield = 57.5)  compared with Poor Husbandry (Low TSW) with fertilizer plus Hybrid seed (yield = 42.5) The profitabilities of these 4 combinations are more dramatically portrayed  in the following diagram:-

The Diamond starts off with all factors at the low level, giving a yield (21.1) which was well above the national average then. by improving all factors, yields were multiplied fourfold. The really important lesson is that yields can be improved considerably by better husbandry without high cash costs; and conversely, it is wasteful to put expensive fertilizers on poorly grown maize.This Maize Diamond concept was successfully used in teaching and for field demonstrations throughout Kenya, and beyond. These  DHTs were a valuable bridge between experimental results and field conditions, and also helped in the interpretation of past as well as subsequent trials results. Time of planting ToP was obviously very important; in 27 other trials over 4 years, each day's delay after the start of the rains reduced yields by 70 kg/ha. ToP also affected N responses; later plantings gave progressively smaller responses. Trials to identify the causes of the ToP effect eliminated various possibilities, including moisture shortages, fungal  build-ups and N shortages on later plantings. Trials were conducted to study the effects of varying amounts of water applied during the first 5 weeks of growth, with 4ToPs. Increasing amounts of moisture during these early growth stages steadily reduced dry matter produced at 5 weeks, and grain yields similarly. Increases in moisture caused reductions in kernel numbers initiated per plant. Follow-on research in an ODA project at Kitale showed that soil temperatures steadily declined after the rains began, and that early growth, kernel initiation and grain yields were highly correlated with soil temperature during these first 5 weeks. Many other agronomic trials were carried out. Fertilizer trials (with good husbandry) established optimum rates for different maize prices and fertilizer costs. Sulphur deficiencies were identified, the causes ascertained and remedies worked out. Optimum plant populations, inter-row spacings and planting techniques were ratified. Weed control methods including herbicides were evaluated. All new information from the field trials was continuously incorporated into the extension efforts, and was also invaluable in developing the national maize policy.