4 o2 DISCOVERY REPORTS 



Throughout the Siphonophora, palpons show a perplexing array of possible functions. In Apolemia 

 (Willem, 1894) and Forskalia (Neppi, 1921) they seem able to eject waste matter through a terminal 

 orifice. In Physophora, however, they are sensitive, prehensile effector organs (Totton, 1954). In 

 Physalia, their precise function is not known, but it seems probable that they are concerned with the 

 feeding or defence of the gonodendra, particularly if the latter drop away into deep water, as Steche 

 (1907) suggests. Alternatively, they may be vestigial structures. 



Only two clearly distinct cell types have been found in the endoderm in regions other than the palpons 

 and gastrozooids. The first of the two types is a gland cell (Text-fig. 2,gl.eti) ; its nucleus is a thin curved 

 disk pushed away to one side of the cell by the pressure of the secretory mass. The secretion apparently 

 consists not of granules or globules (although fixation may sometimes give the cytoplasm a reticular 

 appearance), but of a homogeneous substance. It is P.A.S. positive. In whole pieces of codon- 

 endoderm fixed in F.W.A. and stained in iron haematoxylin a clear area, evidently an opening, can be 

 seen at the free surface of each of these gland cells. The gland cells usually occur singly, and mitosis 

 has not been observed in them. Their function is not known for certain, but from their structure it 

 would seem reasonable to suppose that they are mucus-secreting ('goblet') cells. They very closely 

 resemble goblet cells of the type shown in fig. 64 of Hertwig (1895). 



The second cell-type from the general endoderm shows no secretory inclusions and no digestive 

 vacuoles. It is probably simply absorptive. Mitosis takes place in these cells with approximately the 

 same frequency as in the ectoderm; the descriptions of polysomaty, the binucleate condition, and 

 nuclear fragmentation under stress, etc. (see under ' Muscle ') apply to both ecto- and endoderm. 



In the spadix of the gonophores, the endoderm cells are exceptionally tall and columnar ; they show 

 a multinucleate condition, the nuclei undergoing amitosis (Perez, 1929). The significance of this is not 

 clear. In the ampullae, the endoderm also shows regional specialization, the cells along one side being 

 very small and closely packed together; again, we cannot suggest an explanation. 



The probable sequence of events in feeding and digestion may be reconstructed as follows. On 

 making contact with a fish, the gastrozooids apply themselves to it, spreading out the whole buccal 

 region as a flat disk. The buccal gland-cells then discharge their contents (? proteolytic enzymes), 

 whose powerful corrosive action dissolves away the surface of the fish. Detached particles of food and 

 dissolved substances are then swept back by the rows of long and short flagella towards the basal 

 region of the zooid, where they accumulate. After an unknown time the gastrozooids detach, their 

 mouths close, and the matter in the enteron is subjected to further digestion, under the action of 

 secretions from cells of the buccal type, and from another type of cell located in the basal region. At 

 the same time, solid particles of food are engulfed phagocytically by the cells of the villi, and digested 

 in vacuoles. Further steps in the process are less clear. The gastrozooids have a valve-like structure 

 near the base (Text-fig. 4c) and this is presumably shut during digestion. After a time, it would 

 probably open, allowing the digested matter to disperse throughout the enteron, where dissolved 

 substances would be absorbed by the general endoderm. The bulk of the solid matter which resists 

 extracellular digestion in the enteron of the gastrozooids would be dealt with intracellularly by the 

 cells of the villi. Whether the products of digestion in the villi are set free again, or are used locally we 

 do not know. 



Ejection of waste matter has not been observed, and nothing is known about the time necessary for 

 various stages of digestion nor about the variations in pH during digestion. If a further investigation 

 should become possible, feeding experiments with pH indicators and with substances in suspension 

 like carmine, indian ink, etc., would be of value in solving many of these problems. 



