ALCOHOLISM AND BEHAVIOR OF WHITE RATS 221 



2. Comparison of the test and coiitrol rats 07i the basis of time 



a. Time: averages per different groups of trials. The time data 

 give the number of seconds taken in passing from the entrance 

 of the maze to the doorway of the central food compartment. 

 The observer sat with eyes on the record sheet within the dark 

 box and the left hand on the stop-watch, which was fixed in 

 position on the table by its holder; when the rat was observed 

 to enter the maze the watch was started and it was stopped when 

 the rat entered the final door. Of the various criteria, time is 

 probably most free from errors in observation; the beginning 

 and end of a trial were clearly marked events that could be 

 measured accurately within the limits of the time of the observer's 

 response to a sensory image. 



The first table is based on the average time per trial of each 

 rat for the following groups of trials: first half of training (12 

 trials), second half of traini?ig (12 trials), omitting the first day 

 (21 trials), all training (24 trials), retention (12 trials), and 

 training and retention (36 trials). The first division of the table 

 gives the averages for the group of rats that were given full 

 training, 'completes'; in the second division the averages include 

 the rats that were not given full training, 'incompletes' as well 

 as the 'completes'; the third division of the table includes the 

 'failures,' the 'completes' and the 'incompletes.' Each number 

 in the body of the table is an average per trial in seconds of the 

 averages for each rat ; the number of rats involved in each average 

 is given. Since there is a great difference in the time taken 

 in different parts of the training, the tests and controls are 

 compared by means of ratios instead of by differences. In every 

 case the larger of the' pair of averages has been divided by the 

 smaller number, thus always giving ratios above 1.000. When 

 the test average is larger, that is, the tests taking more time, 

 the ratio is called 'plus' ; and when the control average is larger, 

 that is, the controls taking more time than the tests, the ratio 

 is called 'minus.' Accordingly, a ratio of (+) 1.500 indicates 

 that the test average is one and a half times as large as the 

 comparable control average; the tests took half as long again 



