56 BULLETIN OF THE 



The Types of Ommatidia. 



With the conclusions arrived at in the foregoing account as a basis, 

 an ommatidiuni can be constructed which will serve as a type for the 

 ommatidia of all Decapods. Omitting tlie accessory pigment-cells, this 

 typical ommatidium would be composed of sixteen cells as follows : 

 two cells in the corneal hypodermis, four cone-cells, two distal retinula?, 

 and eight proximal retinulic. In a similar way a typical ommatidium 

 can be constructed for tlie Schizopods. This typo would differ from 

 that proposed for the Decapods, in that its crystalline cone would be 

 formed of two, instead of four cells. 



The ommatidia of the Schizopods and Decapods, as the development 

 of the lobster shows, are closely related. The arrangement of the om- 

 matidia by which hexagonal facets are produced is permanent in Mysis, 

 and temporary in the lobster. The distal retinula) are grouped in the 

 adult Mysis in a way which is reproduced only in the early stages of 

 the lobster. In Mysis the outline Of the nuclei in the corneal hypoder- 

 mis, as seen in sections tangential to the retina, is strikingly crescen- 

 tic. This form is temporarily assumed by the corresponding nuclei in 

 the young lobster (Fig. 5.3). Thus it is evident that the ommatidia of 

 the lobster pass through a stage in w;hich they closely resemble the 

 permanent condition of the ommatidia in Mysis. For this reason, I 

 believe that the ommatidium of Mysis represents a type ancestral to 

 that of the lobster. 



The only important difference which exists between the ommatidium 

 of Mysis and that of the lobster in its early stages is that in the latter 

 the cone consists of four cells, while in Mysis it is composed of only two. 

 If one admits that the ommatidium of Mysis is the forerunner of that 

 of the lobster, this difference in the number of cone-cells can easily be 

 explained on the supposition that the cells which form the cone in 

 Decapods divided once more than those in Hchizopods. If these two 

 types of ommatidia are thus related, it is only natural to expect that 

 this process of cell-division may connect the ommatidium of Mysis 

 with that of some lower Crustacean. 



The ommatidia in the eyes of Amphipods are constructed upon a 

 type which seems thus connected with that of Mysis. In Gammarus, 

 for instance, I have found that the ommatidia possess an undifferen- 

 tiated corneal hypodermis, a ciystalline cone of two cells, and five retin- 

 ulfc. If one desires to convei't this type into that of Mysis, or through 

 i\h-sis into that of the Decapods, the necessary change merely involves 



