The greatest danger in work of this kind is the tendency to generalize and oversimplify the situation. For example, in three carefully controlled feeding experiments carried out at the Plant, Soil, and Nutrition Laboratory, S. E. Smith found that in only one was there any indication of an advantage in favor of the fertilized forage.

The work at Missouri has also indicated a variability in results. Thus, in one of the experiments there, the difference in the average weight of sheep that were fed hay from fertilized soils and those fed hay from unfertilized soils was very slight. The daily average gain in the two was 0.1408 and 0.1644 pound per head. Data are not presented to show that this small difference is significant. In fact, allowing for a reasonable variability in the response of the animals, it can be shown that the greater gains could just as well have been made by either group of the sheep if both had been fed identical hays. In view of the additional fact that 50 percent of the hay from unlimed soils and 46 percent of the treated hay was refused by the animals in this experiment, it is manifestly impossible to conclude that the fertilization or liming program had materially altered the nutritive quality of the forage plant. These instances are cited to show how difficult it is to carry out and interpret these over-all bio-assays of nutritional quality of plants.

It has become obvious lately that soils do not tell the whole story with respect to the nutritive quality of food.

The various factors affecting the mineral content of a mixed hay.

There is the question of sunlight, for instance. Today we know that sunlight is an important factor in the elaboration of vitamins in both plant and animal bodies. The variations in ascorbic acid (vitamin C) in a food, such as the tomato, are amazing. Studies of the causes of such variations have been made by Karl C. Hamner and his associates at the Ithaca laboratory. They concluded that the causes are not so much associated with soil conditions or cultural practices as they are with location. They found that tomatoes of the same variety grown in some parts of the country have three times more ascorbic acid than those grown in another part. Drastic soil treatment failed to modify the vitamin content of tomatoes grown under similar climatic conditions. An explanation for a large part of the variations was found in the amount of light received by the plant. In fact, the effect of variation in the amount of light on a plant is so pronounced that any effect of soil fertility or of fertilizer treatment may be completely obscured. Failure to recognize this point has led to several questionable interpretations in the recent literature of the relationship between fertilization and the vitamin content of foods.

Even the actual amount of sunlight received by the entire plant may not be the deciding factor in determining the vitamin C content of the portion used for food. For example, later and more refined experiments by George F. Somers and Hamner have shown that the vitamin C content of the tomato itself is apparently related to the amount of sunshine that strikes directly on the fruit, for sunshine on the vines apparently has little or no influence on the fruit. Consequently, it is clear that shading of the fruit by an abundant foliage will have an important effect in limiting the vitamin C content. Furthermore, fruit taken from the shady side of the tree or vine or fruit subjected to any other interference from sunlight may be lower in vitamin C than the fruit obtaining direct sunlight. Thus, the total amount of illumination on a field of tomatoes may not necessarily be related to the vitamin content of the fruit from that field because of other factors that are also operating.