The coffee plant does not thrive well under shade in most of the coffee areas of the State of São Paulo, Brazil. It has been shown that water competition from the shade trees during the dry season is the factor responsible for the failure in growing coffee under shade in São Paulo. In all cases where the coffee plants were not thriving well under shade, the soil moisture after two to three months of drought was at the wilting point at the depths most utilised by the root systems of shade trees and coffee plants. In open plantations this was never observed. All commercial coffee plantations of Central America are shaded. Most of them are located in areas with a dry season as long as in São Paulo and some times even longer. A comparative study of the climatological factors, methods of cultivation and most common species of shade trees was made. No one of these factors explained the different behavior of the coffee plant under shade in Central America and in São Paulo. The available soil water was measured in many coffee plantations in Costa Rica and El Salvador after four months without apreeiable rainfall. In all cases it was' found that, the actual moisture percentage was considerably higher than the wilting point. The amount of available water held by soils in São Paulo and Central America does not seem to be significantly different, although this was not carefully determined. It is suggested that a comparative study of the moisture-tension curves of soils of São Paulo and Central America might explain the different behavior of the coffee plant with shade under these two different conditions. It is possible that, as a consequence of a much steeper moisture-tension curve, the plants in São Paulo soils transpire freely until the soil moisture reaches the wilting point. This would result in a rapid soil water consumption. If the Central American soils have a smoother moisture-tension curve, the Avater absorption and transpiration rate by plants may be slower and as a consequence the available water may last longer in the soil. The ratio of the moisture equivalent to the wilting point was found to be around 1.29 for the soils studied in the Meseta Central of Costa Rica, around 1.44 for those in El Salvador and 1.94 for the ashy soils in the vicinity of San Salvador volcano.

The present paper discusses the amount of water lost by the soil through the transpiration of coffee plants and shade trees in shaded coffee plantation. All methods and techniques employed in the study of transpiration of the shade tree Inga edulis Mart, were exactly the same as those described in a paper recently published (4). It was determined previously (3) that a shaded coffee plant receiving 55% of full sunlight transpired 80% as much as when in 100% full sunlight. The amount of water transpired by a coffee plant was calculated by taking 80% of the amount transpired in unshaded plantation (4). Comparing the water lost by transpiration of shaded coffee plants and shade trees with the average rainfall in Campinas, São Paulo, Brazil, it is seen that from April to September the amount of water transpired exceeds the average rainfall. This finding seems to agree with our previous conclusion (1, 2, 5) that water competition between shade trees and coffee plants is the factor that controls the success of growing coffee under shade in the State of São Paulo, Brazil. The highest daily rate of transpiration encountered, for Inga edvlis was 36.7 milligrams per square decimeter per minute. The average daily transpiration was found to be 9.90 grams per square decimeter per day. The total amount of water transpired by an Inga tree was calculated to be 71,140 liters per plant per year. Since the plants submited to the experiment did not suffer a water shortage at any time, the present results show the approximate amount of water lost by the Inga tree and shaded coffee plant under optimal soil-water conditions.

The development of the root system on five varieties of rice grown under upland conditions, has been studied. It has been found, with the rice varieties used and under the conditions tested, that 66 percent of the roots occured within 5 centimeters and 95 percent within 15 centimeters of the surface of the ground. A few roots were found to reach a depth of 1 meter, and occasionally to a depth of 1.35 meter. The Iguape-catete variety of rice showed a tendency to develop a somewhat deeper root system than the other varieties studied. Recognition of the fact that development of the roots of the rice plant is largely within the plowed layer of soil, suggests that deep plowing prior to planting with frequent and shallow cultivation following planting, to eliminate weed competition, would be desirable.

In determinations made with a porometer it was verified that the stomatal opening of coffee leaves increases in size immediately after the leaves are detached from the plant (fig. 1). This fact indicates that the method of rapid weighings of detached leaves is not suitable for studies on the transpiration of the coffee plant. The transpiration of coffee plants has been determined by the weighing potted plant method. Pots with a capacity for 51 liters were used. They were made of galvanized iron plate and had no drainage hole. When the moisture content of the soil approached the wilting point water was added to bring all the soil mass to fiel capacity. In order to avoid oxygen depletion the pots were not sealed. To decrease the direct evaporation at the soil surface a two-centimeter layer of rice hulls was used as mulch, and a celotex cover was placed on top of the pot. Three similar pots with the same type of protection, but without plants were used to measure the direct evaporation from the soil surface. The weight losses of these pots were subtracted from the weight losses of the pots with plants. Two to throe-year old plants of Coffea arabica var. bourbon were used. Three plants were employed each month and after this time their leaves were picked off, and the total leaf area was determined. The tests were carried out for a whole year. The transpiration rates were calculated in g/dm²/day. The total leaf area of an adult coffee plant was found to be 3,146 dm² (average of seven trees). Based on this average and on the transpiration rate of the experimental plants it was possible to calculate approximately the amount of water that an adult coffee plant would transpire under the same conditions. Since the plants submitted to the experiment did not suffer water shortage at any time, the present results show the approximate amount of water lost by the coffee plant under optimal soil-water conditions. The data are presented in table 1. The figure 1 shows the monthly amount of water lost by transpiration in a coffee* plantation compared with the monthly rainfall during a year. The environmental factors that influence plant transpiration vary much less between the same months of different years than between the different months of the same year. Hence the results obtained should not be far from the average; that would be obtained if the transpiration had been measured during a number of years. Comparing the water lost by transpiration of coffee trees with the average rainfall in'Campinas, State of São Paulo, Brazil, it is seen that from October to March rainfall exceeds water consumption. In April, June, and September the excess is small, and in May, July, and August the amount of water transpired exceeds the average rainfall (fig. 2). Since the rainfall distribution is variable, this figure may be different in a given year. The highest daily rate of transpiration encountered was 17.6 mg/dm²/m (over twelve; hours). Previous results obtained by the writers showed 18.2 mg/dm²/m to be the rate of transpiration in a very sunny and hot day. Nutman working in Africa (2) found a maximum daily rate of 20.67/dm²/m (over nine hours). These data are comparable and must represent approximately the maximum rate of transpiration of the coffee plant. The average daily transpiration encountered was 6.20 g/dm²/day and the total amount of water transpired by a coffee tree in one year was found to be 7,273 liters.

Using the method of Briggs and Shantz, several types of soils of the State of São Paulo were studied respecting their wilting points. The results obtained are given in the Table I, and show us that we cannot use as wilting point twice the value of the hygroscopicity as given by Camargo and Vageler for the soils of the State of São Paulo.

In order to investigate if the Citrus disease Tristeza could be caused by a change in the rainfall, its distribution or in air humidity a study of these factors was made based on data obtained from 1890 to 1945 at Campinas, São Paulo, Brazil. A large variation in the data was found but no tendency could be noticed as to increase or decrease of any of factors analysed. This does not support the general belief that climate has been changing and the disease Tristeza cannot be assumed to be caused by change of any of the factors analysed.