In a continuation of an investigation of the constituents of the Porifera for the purposes of chemotaxonomy (I), we have examined the sponge Thoraandra excavatus (Ridley) (order Dictyoceratida).The nomenclature applied to this sponge requires clarification because it represents a new combination in the literature. The species was originally assigned by Ridley to the genus Stelospongus. This was a new generic name introduced as a spelling correction (of the earlier Stelospongos Schmidt), and it is, thus, a synonym of Stelospongos. Stelospongos is unrecognizable (2), and, consequently, any species assigned to the genus, other than the type species Stelospongusfiiabiiis Hyatt, must be reallocated. Stelospongus excavatus belongs in the genus Thorectandra Lendenfeld. The original specimen of this species, figured by Ridley, is in the British Museum (BMNH 82.2.23.211), and this specimen from the Arafura Sea is there designated as holotype. One other specimen referred to in Ridley's original description (3) was from Queensland and was interpreted by Ridley to be a juvenile. This small specimen is not T . excavatus but is Dysidea hwbaceu . Chromatography of the light petroleum ether extract of one freeze-dried sponge gave a 0.07% yield of manoalide 25-monoacetate 111, whereas similar treatment of a second freeze-dried sponge gave a 0.15% yield of thorec-'For Part 11, see R.C. Cambie, P . A . Craw, P . R . Bergquist, and P. Karuso, J . Nat. Prod., 50, 948 (1987).tolide 25-monoacetate [4] and none of compound 1. Extractions of several other individual sponges yielded mixtures of compounds 1 and 4 . A similar variation of related constituents has recently been observed by Faulkner et al. (4) for the sponge Luffclriella variabilis (order Dictyoceratida).Manoalide monoacetate 111 crystallized from CH2C12/hexanes as flakes, mp 117-1 19", [~J D i-34". Examination of the spectral data of this compound established that it was closely related to the pharmacologically active (5-7) sesterterpenoid manoalide 121 that has been isolated (8) from the related sponge L. variabilis. A molecular formula of C27H3806 for compound 1 was suggested by a weak molecular ion peak at mlz 458.2672 in the mass spectrum. The base peak at mlz 137 and fragment ions at mlz 121, 107, and 95 derived from the mlz 137 ion were diagnostic for the presence of a C,,H,, [(trimethylcyclohexenyl)methyl} end group. The chemical shifts of the C ,,H end group in the 13C-nmr spectrum of compound 1 were in good agreement with those of the corresponding resonances for this subunit in the reported spectra of manoalide {2} (8) and also of hydroxymokupalide 151 (9). Furthermore, comparison with the chemical shifts of the aacetyloxybutenolide terminus in the 13C-nmr spectrum of the acetate of mokupalide 161 (9) established that compound 1 corresponded to the monoacetate of manoalide in which the a-hydroxybutenolide rather than the 6-lactol moiety was acetylated (Table 1).
