In the second division, (_b_), of our experiments come those in which a glass screen was placed over the subject's hands. For the first four of these we used a framed window pane which happened to be handy. Then we obtained and used a sheet of 32 oz. glass, measuring 22 by 10 inches and 1/6 inch in thickness. This was supported on two large books placed beyond the subject's hands on each side, and in this position the upper surface of the glass was 2-1/4 inches above the surface of the table, so that there was ample room for the hands to rest underneath without touching the glass. Mr. Smith held his hand in the usual position over the selected finger, above the glass and not touching it. Under these conditions we tried 21 experiments with P., of which 18 were successful, and 6 with Miss B., all successful. In the case of the 3 failures with P., no effect was produced on any finger. In one successful case, the time taken was long, and we interrupted the experiment by premature testing in the way explained above. Division (_c_) includes those experiments in which Mr. Smith did not approximate his hand to that of the subject at all, but merely looked at the selected finger from some place in the same room as the subject, but out of his sight. The distances between him and the subject varied from about 2-1/2 to about 12 feet. Under these conditions we tried 37 experiments with P., 18 in 1890, of which 6 were failures, and 2 only partially successful, and 19 in 1892, of which 10 were failures. The proportion of success was, it thus appears, much less than under the previously described conditions, but still much beyond what chance would produce. Of the 6 failures in 1890, one was a case in which Mr. Smith made a mistake as to which finger we had selected, but succeeded with the one he thought of. In another case the left thumb instead of the right thumb became insensitive. In the other 4 cases no finger at all was affected. Of the 10 failures in 1892, no effect was produced in 4 cases; in another the right (viz., the little) finger of the wrong hand became insensitive;[56] in 4 cases an adjoining finger was affected--once only slightly--instead of that selected, and in the remaining case a finger distant from the selected one was slightly affected.