COLE. — IMAGE-FORMING POWERS OF VARIOUS TYPES OF EYES. 855 



left. The worm is in all five cases headed away from the observer. 

 It will be noticed that the circles on this printed blank are divided 

 into divisions of 10° each, corresponding to the divisions of the circles 

 on the table, so that it was an easy matter to plot the path of the 

 crawling worm with considerable accuracy. The successive trials were 

 numbered in the order in which they were made, from one to five. 

 When the animals were headed in the opposite direction (toward the 

 observer), similar blanks were used, but the zero point of the transverse, 

 or normal, axis of the circles was the near end, not the far end. By 

 this device the relation of the lights to the diagram is kept always the 

 same ; the large light being at the left, the small light at the right. 



In giving numerical value to the angles at which the outer circle 

 was crossed, the readings were, for convenience, grouped into as many 

 classes as there were divisions of the circle. Each class was designated 

 by one of the numbers, to IS, and embraced all the records falling 

 within 5° of the radius bearing the corresponding number. Its nature 

 was further indicated by Lg or Sm, according as the radius in question 

 lay on the side of the normal axis toward the large light or toward 

 the small one. The dividing lines (radii) shown on the diagram there- 

 fore fall in the middle of their respective classes ; thus, for example, 

 the radius numbered 2, which marks a point 20° from the normal axis, 

 is in the middle of class Lg 2, or Sm 2, as the case may be, and all cross- 

 ings between 15° and 25° fall in this class. Crossings which happen 

 to lie exactly midway between the lines numbered on the diagram are 

 always put into the class which lies next to the right of the point of 

 crossing ; in other words, in the direction toward which the hands of 

 a clock move. Thus a reading falling midway between 5 (50°) and 

 4 (40°) on the Lg side of the vertical would be put into class Lg 4f, if 

 it fell in a corresponding position on the Sm side of the vertical, it 

 would be included in class Sm 5. When to the readings for an animal 

 headed away from the observer are added those for the same animal 

 headed in the opposite direction, — the latter would lie in the lower 

 half of the circle, — the possible error resulting from this method of 

 recording is counterbalanced and the effect thus eliminated. 



From what has been said it can be seen at a glance that the five 

 readings shown in the illustrative figure (Figure 5) fell into the follow- 

 ing classes : one in class ; one in class Sm 1 ; one in class Sm 3 ; 

 and two in class Sm 5. 



Something must be said with regard to the disposition to be made 

 of those cases in which the worm turned more than 90°, and so crossed 

 the circle at some point in the dark half of the field. There are three 

 possible ways in which these cases might be treated. Either (1) ig- 



