several species. On the \entral side the\', usualh-, arc- soniewiiat more dewlopfd than on ihc dorsal sitle, whilst in some 

 places, e. g. the dorsal side of the bod\--niargiii. they almost disappear. I have sometimes found the protoplasm of these 

 cells to be striated, as described 1)\- Heard, but tiiat the ciliie are »directl)' connected with the nucleus»: by means of thesi' 

 »threads V is, I think, \-ery problematical imleed. Inside of the epidermic cells, especially on the ventral side of M. giganteum, 

 there are, frequently, larger cells seen containing large nuclei antl nucleoli. From these cells, a prolongation ma\', frequentl)', 

 be seen issuing towards the epidermic cells; whether these prolongations are connected with the epidermic cells, consequently 

 nerve-cells, or if all of them penetrate through to the surface, I ha\e been unable to determine. Some ot them are, in an\' 

 case, gland-cells, because I have distinctly detected excretorj' openings in them. Several such large cells are found situated 

 among the epidermic cells. In the margin of the Ijody, I have found these cells to be in correspondence with nerve-hbres. 

 I have observed nothing resembling the sense-organs, or chitinous hollow rods, described by Beard. 3) The cutis varies in 

 thickness; it is especiall)' thick upon the dorsum, in most species e. g. M. glabrum, M. carpenteri. In M. giganteum it is, 

 on the contrar)', inore developed on the ventrum. I have, also, observed here, in the cutis, unipolar nerve-cells whose pro- 

 longations are directed towards the ventral muscular layer. 4) The ectodermal muscular layer is situated beneath the cutis. 

 Its development varies considerably in the diflerent species, and it appears as if it, usualh', was more developed in the stouter 

 less migrator}- species. I have not obser\'ed such a regular arrangement of concentric muscular rings and raili.al fibres as is 

 described b\' Graft' in M. cerriferum. 



Nervous System. 



This is, in the M}'zostoma, quite in accordance with the usual t)-pe, common to Annelids and Arthropods. It con- 

 sists ot an oesophageal ring with cerebral ganglia, (connected with this ring, there is a proboscideal — stomato-gastric — 

 nerve-sj'stem) and a short ventral nerve-cord. It, altogether, appears to be greatly difterentiated, and the central parts have 

 in a great degree -gagne la profondeur«. A thick muscular layer separates the ventral cord from the ectoderm. 



The oesophageal ring^) with the cerebral ganglia, is, relatively, only slighth' developed; the ganglia have no 

 special sheath but lie entrenched in the connective tissue. The ring-commissures are, on the contrary, provided with a double 

 ncurilemma-sheath, and are connected with the \'entral cord at the anterior extremit)-, between the first pair of peripheric 

 nerves (PI. I, fig. 9, o.cm.). In their inferior part, the_\- are rather slender, and are not connected with ganglionic cells; these 

 parts are ver\' long. In the superior part, the_\- are, on the contrary, usual!)-, surrounded bj- groups of ganglionic cells whilst 

 there are, also, two small ganglia w-ithin the neurilemma-sheath one on each side, which contain a few cells (PI. Ill, fig. 6, a., 

 8, PI. I, fig. 9, a.:. The surrounding cells, usually, occur, as well, in front of as to the Ijack of the oesophageal ring. 

 The}- are arranged in different groups. \ide PI. Ill, fig. 5 — 7 which show 3 sections out of a series taken from this 

 part. In the posterior part of the proboscis, cells are found placed nearer towards the ventral side, where still further back 

 the\- unite so as to form a cellular ring round the proboscis independentl}- of the oesophageal commissures (PI. I, fig. g, s. c. 

 PI. Ill, fig. 7). From the commissures, a pair of nerves issue (PI. I, fig. 9, n", PI. Ill, fig, 6, n.) towards the ventral side, their 

 purpose being, I think, to bring the ventral cells into correspondence with the commissures. I ha\'e found multipolar cells 

 in the posterior parts of the cellular ring (PL III, fig. 10), which lateralh- and dorsalh' is more developed and volu- 

 minous; this part must be considered, therefore, to be analogous to the usual cerebral ganglia of the Annelids, and the 

 ventral parts to be originall}- sprung from the dorsal cerebral part. The prolongations of the multipolar cells in the posterior 

 cerebral parts of the cellular ring are thrected towards the oesophageal and gastric epithelium. I am, therefore, tlisposed 

 to consider this portion of the cellular ring as s}-mpathetic. Several nerves issue towards the gastric epithelium (PI. I\', fig. 

 19, n., fig. 23, n., n.'). One pair of nerves passes posteriori}- along the stomach (PI. I, fig. 9, n.s; PI. I\', fig, i& 16, n.s). 

 I consider these nerves to be nervi sympathici. I have not examined their further course, as I have not been able to deter- 

 mine how the}- are connected with the oesophageal collar. 



The prohofycidcnl nerve system is well developed. Towards the anterior extremity of the proboscis, 3 pairs of 

 nerves issue from the oesophageal ring (PI. I, fig. 9, cm.', cm.", cm."":. These nerves connect the oesophageal ring 

 with the tentacular nerve-ring situated in the proboscis, in front of the muscular bulb. From a series of sections, it can be 

 seen th.it there are m.ui}- nerve-cells situated between the oesophageal ring and the tentacul.ir ring \'ide I'l. Ill, fig. 3 — 6. 

 The cells, as the}- approach this ring, become considerabl}- reduced in number, .ind round about the ring there are, usualh', 

 extremel}- few — if any at all — to be found. These cells communicate, b}- means of minute nerve-branches, w-jth the 3 pairs 

 of longitudinal nerves around which the}- are especiall}- grouped. The tentacular nerve-ring is pro\-ided with a neurilemma- 

 sheath, and inside of this I found no ganglionic cells. I-'rom this ring, nerves issue, one to each of the tentacles, also a great 

 man}- to the anterior part of the prolioscideal canal; ,ind fiuther, four stouter oesophageal nerves issue from it, passing 



M The description of a ring gl>-en by GralT is not correct, such .1 ring docs not really exist, .is may be gathered from the above remarks. 



