(e\v unmistakable species, the determination of alcoholic material cannot be regarded as 

 infallible, and it really seems as if many of the earlier records of Chaetognatha should, 

 properly speaking, be neglected for faunistic purposes, when based solely on alcoholic specimens. 

 Where species are strongly alike, — for example, c?iflata pulchra and (sometimes) hexaptera, 

 robusta and fcrox, Bedoti and Sibogac, neglecta and young serratodentata and Bcdoti, — 

 absolute certainty in such material is often unattainable. In small specimens even close 

 microscopic examination often fails-, in dealing with small (and in most cases young) specimens 

 we are on most uncertain ground, even in finely preserved material ; it seems sometimes as if 

 they might belong to almost any species, for the proportions of tail to body, and the relative 

 proportions of the fins, often alter with age; and of the three parts of the armature — jaws, 

 anterior teeth, and posterior teeth, — the one may increase in numbers more rapidly than the 

 others; again, probably the range of variation, and certainly the probable error in measurement, 

 are greater in young specimens than in older and larger examples. Not even alcohol specimens 

 from the Zoological Station at Naples, prepared with all the care and skill for which that 

 institution is famous, can compare with formalin specimens preserved in the hurry which is 

 inseparable from work at sea. For all plankton except Ctenophora and large fishes, formalin 

 remains bv far the best reagent, if followed in the case of some calcareous organisms by a 

 transference to alcohol. 



The only safe procedure, therefore, in the case of the spirit material of the Siboga, 

 seemed to be to ascertain which of the species represented in the formalin material occurred 

 also in the remaining hauls; in dealing with these one felt on fairly firm ground. -Some deep 

 hauls with the Vertical Net, in which the material, though alcoholic, was in fairly good condition, 

 were studied in greater detail. But it must not be inferred that of necessity no species occurred 

 in the collection beyond those here recorded; indeed, it is nearly certain that at least one new 

 species was captured which is not referred to below, but a diagnosis based merely on the 

 formula of its armature would be valueless. 



No great attention has been paid to the numbers of specimens of each species 

 captured in the different hauls; because (fortunately for the reporter), not the whole collection, 

 but only picked specimens, were sent. Some idea, however, of the numbers in which a species 

 occurred may be gathered from the numbers given in the general table, (p. 28), taken mainly 

 from the material preserved in formalin ; these are doubtless approximately accurate, although 

 perhaps not absolutely exact, owing to the close resemblance between certain species. 



b. THE CHARACTERS ADOPTED FOR DIAGNOSIS. 



This report follows in general lines the recently published report on Chaetognatha from 

 the Bay of Biscay x ), which will be quoted below for brevity simply as 'Biscayan Plankton', or 

 the 'Biscayan Report'. The introduction thereto, 'On the systematic determination of Chaetognatha', 



1) Biscayan Plankton: collected during a cruise of II. M.S. Research 1900. — Part III. The Chaetognatha: by G. Herbert 

 FOWLER. — Transactions of the Linnean Society of London. Series II, Zoology, Volume X, pp. 55 — 87, 1905. 



