SUMERIAN COPPER—MECHANICAL ABILITY 305 
Ar- 
Tell Asmar, Early Dynastic. Copper. Tin. Nickel, senic.Lead. Iron. 
Dagger blade . 1329 90°82 2°63 0°33 O°15 0 O°I5 
Sickle blade . sb Jgagx -/88"'rTii0 ° 2°23. 0 0°40 
Arrow butt . . 1048 61°50 oO 0736. 0°37 0 — 
Arrow butt . . 1097 70'0 trace 0°39 0°86 
Rolled pin. . 1038 95°49 0 0°30 1:27 10 1°30 
Rolled pin . ’ 986 87:81 0 0:90 2°08 Oo 1°04 
Khafaje, Early Dynastic. 
Dagger blade . 1296 49°82 0 6°29 0°94 022 °° — 
Rolled pin . epAOR. 520570, 10 ° ° trace 
Rolled pin . 1 187 77°98 o trace 0°22 0 
Lump. : 3 152. 78°73 6°31 0°22 0°90 0 
Alishar Héyiik, Anatolia, 
Copper age and transt- 
tion period. 
e—700 58:12 4°82 0°04 0°14 1°22 
e—833 61°36 10°8 o751 0°03 0 
e—o963 60°52 0°76 trace orII 0°40 
+ Much sulphur. 
MECHANICAL ABILITY. 
Final Report of Committee appointed to inquire into the factors involved in 
Mechanical Ability (Dr. C. S. Myzrs, C.B.E., F.R.S., Chairman ; 
Dr. G. H. Miss, Secretary; Prof. C. Burt, Dr. F. M. Earte, 
Dr. Li. Wynn Jones, Prof. T. H. Pear). 
ARISING out of the work reported to the Association by the Committee 
on Vocational Tests in 1931, the present Committee arranged for a thorough 
revision and statistical examination of the results to be made. As an 
outcome of this work the need was apparent for a further examination of 
the routine manual factor which plays an important part in the process of 
repetitive manual operations. 
I. SuITABILITY OF MEASURES. 
Reliability.—The suitability of the measures upon which the conclusions 
of this report are based was carefully investigated. The reliability of the 
‘ mechanical ’ tests (assembling and aptitude), as indicated by their inter- 
correlations, was in the neighbourhood of 0-6 to.o:7. ‘That of the routine 
tests, based on the correlation of one test with a general repetition, varied 
from 0-52 to 0-91. The reliability of the ‘intelligence’ tests, based 
on the correlation between the sum of the odd sub-tests and that of the 
even, varied from 0:8 to 0°9. 
The intercorrelation of each ‘trial’ with the other trials constituting 
the test was examined in the data on the routine tests obtained from the 
adult subjects. The figures indicated that the reliability of manual tests 
depended primarily on the number of repetitions (or ‘ trials ’) rather than 
upon the complexity and length of the ‘trial’ itself. About the same 
degree of ‘ reliability ’ could be expected from a given number of trials, 
M2 
