YELLOW-BROWN CELLS OF CONVOLUTA PARADOXA. 443 
that the yellow-brown cells of C. paradoxa are present in 
the egg. 
Careful examination shows, however, that this view is 
erroneous, and that the orange granules of the egg have 
no relation whatever with the yellow-brown cells of the 
larva or‘adult. The pigment-bodies of the egg have none 
of the histological features of the yellow-brown cells of 
the larva or adult; the pigments are distinct in colour and 
react differently with reagents. Thus, whereas the orange 
granules of the egg and also the orange glands of the animal 
become colourless when treated with 90 per cent. alcohol, 
the yellow-brown cells of the animals become green. In 
other words, these cells contain two pigments, a yellow and a 
green, whilst the orange granules contain only one pigment. 
Further proof that the orange granules are in no way what- 
ever related to the yellow-brown cells is given in Section III. 
(e) Ingestion of solid food by C. paradoxa.—Imme- 
diately after hatching C. paradoxa like C. roscoffensis 
begins to ingest solid bodies (Pl. 26, fig. 4). 
But whereas the latter species, soon after its infection by 
the green cells, ceases altogether from ingesting solid food, 
C. paradoxa continues throughout its life to be a voracious 
feeder. Its staples of diet are diatoms, many dozens of which 
may often be found in all stages of digestion lying in digestive 
vacuoles (PI. 26, fig. 4). In addition to diatoms C. paradoxa 
takes up almost any small organisms it comes across. It 
also swallows and digests relatively enormous copepods ; 
whilst later in the year its body may become so studded with 
the tetraspores of various alge as to give to the animal a 
dark red colour. Failing suitable food-bodies C. paradoxa 
swallow greedily any inert, granular substance supplied to it. 
When supplied with congo red, masses of this pigment are 
ingested, and lie in colourless vacuoles in the solid gut. The 
congo red retains it colour in these vacuoles, indicating that, 
though ingestion does not require any chemical stimulus, the 
secretion of digestive juices into the vacuoles depends on 
such a stimulus. Itis probable that the digestive secretion 
