Weight-accelerating movements:
The average of seven subjects:
The accelerating weight reduces mean and constant errors, and improves regularity of curves, except for l.h.e.c. constant-error record. There is some evidence that a pull causes overruling, while a retarding weight causes underruling, but there are exceptions enough to warrant care in finally accepting this statement. Visual control with accelerating weight reduces error more than the weight acting alone.
For mean errors,
(1) weight reduces errors for r.h.e.o. and l.h.e.c. as compared with free-movement records, while the other two curves are inconclusive;
(2) visual sense helps in accurate ruling;
(3) non-visual records are not reduced, as a rule, from the results of free motion.
For constant errors,
(1) the accelerating weight tends to greater accuracy, with an exception for the-non-visual records.
No testimony of marked importance is to be noted in comparison of right-handed and left-handed subjects; the more trained hand shows greater accuracy in some cases, but fails to excel in others; so the data is inconclusive.