218 The Philippine Journal of Science 1921 
typical series of tests, together with the maximum variation 
from the mean compressive strength. It is apparent that the 
grand mean value of 34 per cent obtained by averaging the 
individual maximum variations from the mean compressive 
strength of the field-made specimens is excessively high and 
could have been reduced considerably by more accurate and care- 
ful field work. 
The maximum variation from the mean compressive strength 
of laboratory-made specimens is between 3 and 20 per cent and 
the grand mean variation is only 9 per cent, which is about one- 
fourth of that shown by the average results obtained from field- 
made specimens. By making smaller batches and giving greater 
attention to detail, the results from test specimens made in the 
laboratory will usually be more concordant. However, with due 
care, equally uniform results can be obtained from specimens - 
made in the field. 
Insufficient mixing that yields a nonuniform concrete; varia- 
tion in the percentage of water; inaccuracies in proportioning 
cement, sand, and gravel; mistakes in marking test specimens; 
variation in storage conditions of the concrete test pieces—are all 
contributing factors to the erratic results already noted. How- 
ever, there is still another factor which considerably affects the 
strength of concrete test specimens; namely, the manner in 
which the concrete is placed in the mold. It is certain that not. 
a little of the variation in results is due to the different methods 
of molding the specimens comprised in a single series. Some 
of the specimens were made without tamping or spading and, 
of course, were honeycombed with air blebs. Others were tamped 
and consequently showed a compact texture. A uniform method 
of molding is conducive to uniform ultimate compressive strength 
but naturally cannot overcome the other disturbing factors which 
are the chief cause of irregular results. The more uniform 
the materials, the more carefully they are mixed; and the more 
uniformly and carefully the concrete is placed in the mold, the 
more uniform will be the ultimate compressive strengths given 
by the test specimens. 
SUMMARY AND CONCLUSIONS 
Ultimate compressive strength tests have been conducted on 
1,677 concrete specimens more or less representative of concrete 
made in the Philippines during the decade from 1908 to 1918. 
Laboratory-made concrete specimens proportioned 1 : 2 : 4 and 
aged 28 to 31 days give an average ultimate compressive strength 
