Several tornariae were obtained from the plankton collected during the Cruise of the "Baependi" and the "Vega" to Trindade Is. and vicinities on 18th June 195D, also from plankton collected at Fernando de Noronha Is. These larvae were fixed in a solution of formalin at 4%. Those of Trindade Is. were in a very bad state of conservation (Fig. 1) . They proved to be probably Tornari a weldoni Stiasny in a young Krohn stage of progressive development (no coeloma were noticed). The secondary loops and lobes which were 4 to 6 in number, the high oral arch, the very large oral field, the elevated and narrow ventral loop are characteristic of the above mentioned species and were observed in our best preserved specimen. It is only 1,01mm long, therefore smaller than the T.weldoni described by Stiasny & Stiasny-Wijnhoff (1996). Distribution:- Bahamas Is. and New-England. The beautiful large tornariae from Fernando de Noronha Is. (4mm in height)(Figs. 2-6) are all T.chierchiai I Stiasny in the Krohn stage of development. This is a Ptychodera type larva, according to Stiasny and Stiasny-Wijnhoff (1926), with lobate tentacles, no anal ring, not apically pointed, with a narrow oral field and very deep superior lobes with an inferior bifurcation (Fig. 2). The oral arch is very narrow and low (Fig. 5). The inferior dorsal lobes are narrow, long and without lobate tentacles (Fig. 4). The lateral lobe is short, oblique (as in T. morgani) , narrow, with more or less 8-10 lobate tentacles and provided with a lateral loop as in T.morgani (Fig. 4 and 2), the connective belt between the median and the lateral loops of the post-oral and the pre-oral fields is provided with 6-10 lobate tentacles (Fig. 2). There are about 26 lobate tentacles on the loops. There are no lobate tentacles on the apical region, where the eyes are situated (Fig. 3). Hydrocoel has large "spurs". There is a strong pigmentation on the ciliated ring. The trunk and collar coeloma are periferically situated. The larva is in an advanced Krohn stage of development. T. chierchiai I has only been found in the Pacific up to now. It is generally considered to be the larva of Ptychodera flava widely distributed in that Ocean. A species very closely related to P.flav a occurs in the Atlantic - Ptychodera bahamensis - whose larva T.morgani, differs from T.chierchia i I chiefly by the absence of the inferior dorsal lobes. The larva found at Fernando de Noronha Is. is T. chierchiai I or a morphologically intermediate type between T. chierchiai I and T. morgani. Its occurrence in the Atlantic near to the West Indies strongly supports Trewawas' (1931) and Van der Horst's (1939) suggestions that P. flava and P. bahamensis are probably varieties of the same species. Probably their larvae are also very variable. Or else, T.morgani and T. chierchiai I are not the larvae of the mentioned species of Ptychodera , which is not probable. T. morgani has been reared up to a young Ptychodera sp. stage. Dr. K. P. Rao in India is now trying to rear the larva of P. flava and thus to settle all this controversy. Several tornariae were also sent to me from Algiers (Mediterranean Sea) by Prof. A.Hollande. They are beautifully coloured and mounted in Canada balsam by Dr. J. Cachon, who collected them on the 28th March 1949. One of them had been reared to a stage of metamorphosis. These tornariae belong to two different species: T.krohn i Stiasny and T.dubi a Spengel, both already known from the Mediterranean Sea. T. krohni is very similar to T.mielcki (Stiasny & Stiasny-Wijnhoff 1926, p. 160-161) and my specimens in the Krohn stage of development were intermediate in shape and characteristics between the two above mentioned species, thus proving that the two are varieties of the same species. The height of the tornariae were from 1,1mm to 1,3mm (smaller than the first T. krohni described, probably in part due to fixation and mounting). They belong to the Balanoglossus and Glossobalanus type of tornaria, with the anal field in the shape of an inverted cone, the anterior part of the larva bell shaped, a median apical plate, a narrow ventral loop, which is not very elevated (Fig. 7). The lateral lobe with its corresponding lateral loops is like T.krohni's from Messina (Fig. 10). The stomach is balloon shaped. The primary and the secondary dorsal lobes are very large (Fig. 8). The secondary loops are narrow and protruding (Fig. 7). The oral arch is elevated (Fig. 7). The hydroporus is situated slightly to the right of the median line (Fig. 8). Strongly developed primary inferior dorsal lobes. There are 4 to 5 secondary loops ventrally (Fig. 7). Dorsally there are 3 to 4 (Fig. 8) of the same. The hydrocoel has no "spurs" and the pericardium is already developed. Both collar and trunk coeloma are present. The tornaria is therefore an advanced Krohn stage of development. Fig. 11 shows a tornaria of this species in the Spengel stage of development (with greatly developed coeloma and hydrocoel). Fig. 12 shows the same species at a very late Metschnikoff stage of development (with only very poorly developed trunk coeloma and the beginning of the secondary loop formation, the hydrocoel with thin walls). T. krohni i Spengel is here considered a different species of tornaria probably belonging to Balanoglossus clavigerus, as stated in Stiasny & Stiasny-Wijnhoff (1926, p. 66). Tornaria dubia Spengel collected in the same region, in the vicinities of Algiers, (Figs. 13, 14, 15 and 16) belongs to the Glandiceps type of Stiasny and Stiasny-Wijnhoff. The tornaria is double-cone shaped, with pre-oral field in the form of an upside down Y, and with protruding not separate lqops which give it also the shape of a butterfly. Oral arch with not parallel limbs. The ventral loop has two inferior ventral lobes and two corresponding loops (Fig. 13). It is not long, but large and flat. No lateral lobe, nor lateral loop (Fig. 15). Dorsal primary loops partly in the ventral region with one or two secondary lobes (resulting in an S shaped band) with a long narrow horizontal connection to the median field (Fig. 15). The hydrocoel is very large with pericardium and coeloma well developed. This and the great height of the tornaria (1, 23-1,34mm) show that the larvae of Figs. 13, 14 and 15 are in an advanced Krohn stage of progressive development. They are more developed than the larvae figured by Stiasny & Stiasny-Wijnhoff (1926, fig. 36a and 36b) and by Spengel (1893, pi . 22, fig. 17, 18, 19). Fig. 16 shows this larva in the Spengel stage of development with smaller number of loops and lobes and coeloma far more developed as well as larger hydrocoel. Two metamorphosis stages are also figured (Figs. 17 and 18) but have not been described because they were already mounted in Canada balsam. They belong to T. dubia or to T. krohni and were also obtained in Algiers. It would be interesting to rear both larvae and to settle the question as to which adult species they belong.
