DOUBLE STRUCTURE IN PLANARIA LUGUBRIS. 135 



determined by that of the rest of the part. If the latter should 

 grow to the normal full size the heads would, no doubt, also be- 

 come full size. Since the parts would become full size if they 

 were separated from each other the explanation of their failure 

 to enlarge beyond a certain size is unquestionably connected with 

 the fact of their union with each other. Very often the pieces 

 partially reunite after the operation, and two proboscides are 

 formed as shown in Fig. 16. In such cases the body at, and 

 beyond, the region of the proboscis is broader than normal, and 

 in consequence the heads at the anterior ends may be larger than 

 when only a single proboscis is present, as in Fig. 14. 



In order to see how important a factor the extent of the sur- 

 face regenerating a head is, as compared with the size of the area 

 of union of the parts, the following experiment was made. The 

 old head was first cut off, and then by an oblique cut, as shown 

 in Fig. 17, a triangular piece was partially cut off. Pieces of 

 this sort tend to unite and must be for a time kept apart. The 

 new head develops on the anterior cut surface of the triangular 

 piece, and the other head on the anterior edge of the oblique cut 

 surface. A new pharynx develops along the posterior edge of 

 the oblique cut in some cases, Fig. 18 and 20, in others not, Fig. 

 19, depending, in part, on the extent to which the pieces are 

 kept apart after the operation, in part on the nearness of the cut 

 to the region of the old pharynx. As soon as the two new 

 heads have been fully formed it is seen that their size bears no 

 relation to the fact that one has developed on the cross-cut sur- 

 face and the other on the oblique surface. Their sizes depend 

 rather on the size of the part from which they have developed, 

 Figs. 1 8, and 19, and w r hether a new pharynx has appeared in the 

 triangular piece, and also on the relation of the part to which 

 they belong to the rest of the worm. The following examples 

 may make this clearer. In the first case, Fig. 18, the triangular 

 piece has rather a broad attachment to the other part and con- 

 tains a pharynx. The larger of the two heads belongs to the 

 old part. In the second case, Fig. 19, the triangular piece is 

 much smaller and ends in a smaller head. It does not contain a 

 pharynx and is attached to the other part by only a narrow area. 



The third case, Fig. 20, shows that the triangular piece is 



