216 E. C. MACDOWELL AND E. M. VICARI 



e. 'Failures,^ 'incompletes^ and 'completes^ 



If the rat did not happen to reach the center in a reasonable 

 time on the very first trial, it was removed, fed in the center, 

 and tried again the next day. In such cases all the time spent 

 on different days in reaching the center for the first time was 

 counted as the first trial. In certain cases (those rats marked 

 'F' in the pedigree chart) after a rat had made one or more 

 successful trials and then failed to reach the center after from 

 ten to twenty minutes, the rat was called a 'failure' rat, or just 

 a 'failure.' The time given before such trials were called off 

 was not always the same, for in some trials the rat would con- 

 tinue to be active and keep investigating the alleys, in other cases 

 the rat would give up the problem and lie still. When the rat 

 did this, it was found useless to wait longer, while it seemed 

 wiser to let a rat continue its searching if it was active, and 

 especially if it was in an inner alley. During the experiment the 

 difficulty of handling the data of rats with such failure records 

 was not fully realized, or else a different plan would have been 

 followed. As it was, the rats with these failures were continued 

 throughout the training (in most cases the rat succeeded on the 

 day following a failure); at the end of their training they gave 

 about as good records as the rats in the same group that did not 

 have failures. However, the influence of such failures on the 

 following days remained a serious question. The trials immedi- 

 ately following are slower than the average of the other rats in 

 the same group; even though this difference seems to disappear 

 in later trials, the effect may none the less be operating. This 

 makes the treatment of the data from such rats a matter of 

 difficulty. The time spent on an unsuccessful trial could be 

 added to the following successful trial as was the practice when 

 a rat did not succeed on the first trial. But failing after the 

 center has once been reached is an entirely different matter 

 from not reaching the center for the first time. These failures 

 are not of frequent occurrence; they never occur after the third 

 day, and most of them occur on the first day; only ten rats out 

 of sixty had a 'failure' and only one rat failed a second time 

 after another successful trial. It seems as though some unusual 



