OVERHANG AT BARBERS HILL, TEXAS 33 
ence in those cases, however, was small between ce and be, 3.15 against 
3.24; 3.43 against 3.50; 2.13 against 2.70; 2.69 against 3.24. 
The relative standing of the various forms in reference to the 
order of occurrence of their respective mean squares is given in 
Tables II and III. In Table II, the ‘‘mean”’ mean squares and the 
median mean squares of the series, I-VIII, are given in order of their 
TABLE II 
Weight Means Weight Medians 
25 1.63 ce 25 1.17 ¢é 
1.65 cd 1.21 cd 
2.246f 2.30cd 
2.29 ce 1.38 cd 
2.39 be 1.44 ce 
20 2.42cd 20 1.44 e 
2.61 cd I.50¢e 
Ae Git Wi 1.50¢cd 
2.80e 1.59¢f 
2.80 ce 1.69 cd 
15 2.85 cd 15 1.69 6f 
2.89 cof 1.90 bf 
2.99 ce 1.92be . 
2.99 be 1.946f 
3.00cd 2.02 cd 
be) 3.26 bf 10 2.03 be 
3.30 cd 2.03 of 
3.37 be 2.13 be 
3.40 6f 2.13 be 
3.46 be 2.17¢d 
5 3.50 be 5 2.18 be 
3.53 cd 2.25 Of 
3-576 2.40 bf 
3.69 ce 2.57 cd 
3-69 bf 2.69 be 
° 3.87 ° 2.70 Ce 
3-90 2.73 of 
3-94 3-246 
3.24 Ce 
3-50ce 
magnitude for the mean squares, less than 4.00 for the means, and 
3.50 for the medians. The relative position of each form in regard to 
the magnitude of its mean square is given in Table II for the first 
25 in each list. First place is ranked 25 and last place, 1. 
Although strictly the least mean square should indicate the most 
probable set of assumptions, there is considerable probable error in 
our calculations and we can not be certain that the least mean square 
of our calculations actually is significantly less than some slightly 
larger mean square. Tables II and III give a study of the distribution 
693 
