250 
Journal of Agricultural Research 
Vol. XXXI, No. 3 
The colloidal silica used in these tests was a silica gel prepared 
chiefly by M. S. Anderson, of the Bureau of Soils, by treating a 
sodium silicate solution with H 2 S0 4 and HC1 until a strongly acid 
reaction persisted. The precipitate was then washed many times 
until the wash water had the same reaction as the distilled water 
used (P h 6.2). It was then dried on the steam bath and ground to 
pass a 1-millimeter sieve. 
A mixed gel containing silica, alumina, iron, and small quantities 
of lime, magnesia, and potash, was also tested in one pot for com¬ 
parison with the pure silica gel. This was prepared by adding 
the chlorides of aluminum, iron, etc., to a sodium silicate solution in 
such proportions that the resulting mixture contained silica and the 
various bases in the same proportions as the average soil colloidal 
matter. The resulting gel was washed until the wash waters were 
free from chlorides. The washing apparently removed a consider¬ 
able part of the lime, magnesia, potash, and soda, and a small part of 
the iron and alumina. 
Golden millet (CJiaetochloa italica) was used as the test crop. 
This responds markedly to phosphoric acid fertilization, and, like 
the other Gramineae, it has a low “feeding power” for insoluble 
phosphates. Glazed earthenware pots of 1 gallon capacity were 
used as containers. In each pot was placed 5,030 grams of sand, 
which was maintained at a moisture content of 8 per cent of the dry 
weight by the addition of distilled water, the pots being weighed 
daily. 
The special additions to the different pots—rock phosphate, acid 
phosphate, and silica gel—were thoroughly mixed with the upper 
half of the sand in the pot prior to planting. The rock phosphate 
used was a Florida pebble rock ground to pass a 100-mesh sieve. 
It contained 31.63 per cent of total phosphoric acid. The acid 
phosphate contained 18.72 per cent of available and 19.74 per cent 
of total phosphoric acid. Both phosphates were applied on the 
basis of the total phosphoric acid. 
Each pot in the experiment received the following salts, applied in 
solution, as a basic fertilization to supply all essential nutrients except 
phosphorus: Ca(N0 3 ) 2 .4H 2 0, 1.070 grams; (NH 4 ) 2 S0 4 , 0.491 grams; 
KC1, 0.446 grams; MgS0 4 .7H 2 0, 0.914 grams; Fe 2 (C 4 H 4 0 6 ) 3 H 2 0, 
0.074 grams. 
Seven millet plants were grown in each pot from July 29 to August 
31, 1923. At the end of 33 days of growth the plants were in the 
joint stage, and those in the pots receiving the largest amounts of 
acid phosphate appeared about as large as could be grown without 
danger of having further growth restricted by size of pot or supply 
of basic nutrients. The pots were kept in a glass house and the order 
of the pots was shifted daily. The appearance of the plants well 
supplied with P 2 0 5 was normal at all times (pi. 1). 
The special treatments of the different pots and the oven-dry 
weight of the plants harvested are shown in Table I. Most of the 
special treatments were triplicated. 
The greenhouse of the division of Soil Fertility Investigations, 
Bureau of Plant Industry, was used for the conduct of this experiment. 
