Two new irritants, lyngbyatoxins B [5] and C [6] were isolated from Lyngbya majuscula collected at Kahala Beach, Oahu, Hawaii, and their structures were elucidated. Lyngbyatoxin A [1] was first isolated (0.02%) by Cardellina et al. (1) from a shallow water variety of the marine bluegreen alga Lyngbya majuscula (Dillwyn) Harvey collected at Kahala Beach, Oahu. This compound was later found to be identical with teleocidin A-1 [2] (2). The structure of teleocidin A-1 was determined on the basis of spectral comparison with teleocidin B [3], the structure of which had been established by X-ray crystallography (3,4). The absolute configuration of 2, and therefore of lyngbyatoxin A [1], was elucidated by comparison of its cd spectrum with that of indolactam V [3] (natural and synthetic) (5) and by chemical degradation (2). In Streptomyces mycelia the metabolite teleocidin A-1 is often accompanied by the C-19 epimer, teleocidin A-2 [4]. However, only the former isomer appears to be present in L. majuscula. It is not presently known whether 1 is produced by the cyanophyte or by a microorganism associated with the alga. One of us (REM) had noted over a decade ago that minor amounts of other lyngbyatoxins were present in the alga. We were therefore prompted to reexamine L. majuscula from Kahala Beach to identify these other congeners and to confirm that 4 is not produced. Extraction was carried out on the freeze-dried material by using 50% i-PrOH/CH₂Cl₂ at room temperature. Even though we could not find 4, two new teleocidin-A-type compounds, lyngbyatoxin B [5] and lyngbyatoxin C [6], were obtained. At the same time five known compounds, malyngamides A and B, majusculamides A and B, and debromoaplysiatoxin, were isolated and characterized in addition to 1. Compound 5, an oily syrup, showed a molecular ion peak at m/z 453.3003 (calcd for C₂₇H₃₉N₃O₃, 453.2991). Its uv spectrum was superimposable on that of 1. A second new compound, lyngbyatoxin C [6], had the same molecular formula, as shown by its high resolution mass spectrum (observed m/z 453.2987). The ¹H-nmr spectrum of 5 indicated the structure depicted. {The numbering employed in this paper is based on the proposed biogenesis of the teleocidin/olivoretin class of natural products (6). For more recent results on the biosynthesis of this type of natural product, see Irie and co-workers (7,8).} It was evident that the skeletal structure consists of an indole moiety and a 9-membered lactam ring. The main difference between 5 and 1 could be seen in the ¹H signals for the monoterpenoidal side chain at C-7. Signals were present for terminal methylene protons on C-27 (br 1H singlets at 4.81 and 4.90 ppm), a methyl attached to a double bond at C-26 (3H singlet at 1.57 ppm), and a hydroxyl-bearing methine proton on C-25 (a broad 1H signal at 3.99 ppm which changed to a triplet with a coupling constant of 6.3 Hz on addition of D₂O) and verified the proposed structure. Further support was obtained from the eims spectrum. The molecular ion peak was the base peak, and the fragment ion arising from the cleavage between C-19 and C-23 was found at m/z 354 and was only moderate in intensity (24%). Other fragmentations were also quite similar to those observed for 1. The eims spectrum of lyngbyatoxin C [6] revealed a molecular ion peak at m/z 453, which was much smaller in intensity (8%) than the one for 5, and a fragment peak at m/z 354, for cleavage between C-19 and C-23, which was the base peak. These data were compatible with the structure shown. The cleavage between C-19 and C-23 is favored because it represents allylic and benzylic cleavages. The nmr spectrum also strongly supported this structure with a 6H singlet at 1.20 ppm for the two methyl groups on C-26 and an unresolved broad singlet at 5.41-5.44 for the two olefinic protons on C-24 and C-25. In our previous paper (2), the absolute configuration at C-19 was determined to be R for 1 and S for 4. The cd spectra of these two compounds (Figure 1) showed the only difference at around 230 nm. This observation was further confirmed by the corresponding synthetic materials (9). The cd spectra of 5 and 6 were essentially the same as that of 1, indicating that the configurations at C-19 are R for both compounds. The stereochemistry of the secondary alcohol at C-25 for 5 and the geometry of the Δ²⁴ double bond for 6 remain to be determined. The 50% inhibition for specific binding of ³H-TPA ([³H]12-O-tetradecanoylphorbol-13-acetate) were ED₅₀=2.2 μM for 5 and ED₅₀=0.2 μM for 6, which correspond to 1/200 and 1/20 the activity of 1 (10). Both compounds showed a positive response in the mouse ear irritant test.