284 DR. ROBERTS ON THE DIURNAL 
Before proceeding to discuss the explanation to be 
offered of the power of a meal to depress the acidity of 
the urine, attention must be called to the variation in the 
amount of solid matter—%in other words, solid urine — 
secreted at different hours of the day. The numbers were 
obtained, as already explained, by calculation from the 
density. With the exception of the seven days included 
in Table III., the specific gravity was taken by a hydro- 
meter, and not with that scrupulous care, nor under that 
correction which is required for calculating the solids. 
While therefore able to state that all the tables show re- 
sults remarkably concurrent, I am inclined to place most 
reliance on the determinations in Table III., which were 
made with the necessary precautions. On passing the 
eye down the columus of solids in the six tables of means, 
it is seen that the quantity per hour began to increase 
within an hour after breakfast, that is, between eight and 
nine, and went on increasing rapidly until eleven. It re- 
mained stationary until about noon, and then began to 
fall, and continued to diminish until dinner (see especially 
Tables I., IX. and XIII., when dinner was at four), and 
even for an hour after. At the end of the second hour 
after dinner there was usually, but not always, a slight 
rise. During the third hour there was a very decided rise, 
which went on increasing until the fifth or sixth hour, 
about which time it reached the culminating point. From 
that time the discharge of solids fell gradually until bed- 
time, and sank to the lowest point during the hours of 
sleep. All the tables of means exhibit the same process- 
sion of numbers, whatever the nature of the food might 
be. Exceptional instances occurred, it is true, here and 
there; but with so great a multiplicity of details it was 
impossible altogether to avoid errors, from faulty obser- 
vation, calenlation or record. The contradictory facts, 
however, were quite insignificant beside the overwhelming 
majority of concurrent observations. 
