An examination of the nuclear magnetic resonance spectra of the diacetates of the two isomers reveals that isomer I is 5-nitronuorescein diacetate (Fig. 1) and isomer II is 6-nitrofluorescein diacetate (Fig. 2). We chose the nitrofluorescein diacetates because of their appreciable solubility in chloroform. It was presumed that the oxygen bridge and the O-acetyl groups would shift the protons in the xanthene ring to higher field (with respect to benzene), while both the lactone carbonyl and nitro groups would shift the phthalyl proton peaks to lower field. The spectrum arising from the protons of the phthalyl ring would then be recorded without interference from other aromatic hydrogen resonances, and interpretation of the splitting patterns for the two isomers would allow the structure to be assigned unambiguously.The phthalyl portions of the spectra are shown in Figs. 3 and 4. Each compound had a methyl to aromatic proton ratio of 6 : 9 with the methyl hydrogens absorbing at 139 c/s. For each isomer, six of the aromatic hydrogens resonating at highest field were found in two peaks at 417 and 436 c/s, respectively, in the ratio 4 : 2. The absence of splitting is curious, but these resonances are unquestionably associated with the xanthene ring, because they appear in both spectra. (The expected splitting pattern for the xanthenyl protons is observed in deuterium oxide at a pBL of 8-5.) All spectra were obtained at concentrations of 12-20 per cent (w/v) in deuteriochloroform using tetramethylsilane as an internal standard with a Varian A-60 spectrometer. Inspection of structures 1 and 2 leads to the conclusion that the spectrum arising from the phthalyl ring of the former would be from an ABX proton pattern, while that from 2 would be of an ABC set of hydrogens3. Among the resonances expected in each case would be a peak without ortho splitting, at C4 in 1 and C7 in 2. Because the proton at C4 in 1 has two ortho electron withdrawing groups, while that at C7 in 2 has only the nitro group in an ortho position relative to it, the first order singlet (no ortho splitting) for structure 1 should appear at a higher frequency than that for 2. A singlet appeared at 539 c/s for isomer I and 489 c/s for isomer II, with each showing meta splitting. The spectra of two model compounds, methyl 3-nitrobenzoate and methyl 3-nitroacetophenone, reinforced our interpretation.The chemical shifts and coupling constants enabled the remaining peaks from both isomers to be assigned. Besides the 539 c/s peak, the other member of the AB pair in isomer I (Fig. 3) is the C6 hydrogen, which produced the 518 c/s doublet with fine structure (J6_7 ~ 8 c/s, J6-4 ~ 2 c/s) attributed to meta splitting. Finally, there was a sharp doublet at 450 c/s which, because of the absence of meta splitting as well as its small shift from benzene, fits the X proton at C7 in 1. The assignment of structure 1 to nitrofluorescein I diacetate is further supported by the integration ratio (1:1:1) and the 89 c/s spread for the phthalyl proton resonances caused by the wide variation in electron density at carbons 4, 6 and 7. Fig. 1.Fig. 2. Fig. 3. Phthalyl portion of the nuclear magnetic resonance spectrum of nitrofluorescein I diacetate measured on a 12 per cent solution (w/v)in deuteriochloroform.Fig 4 Phthalyl portion of the nuclear magnetic resonance spectrum of nitrofluorescein II diacetate measured on a 13 per cent solution (w/v) in deuteriochloroform. The five-line (first order ortho splitting) pattern from the phthalyl ring of isomer II (Fig. 4) was spread only by 30 c/s, which indicates that the three protons exist in a similar electronic environment, as in structure 2. The presence or absence of meta splitting permitted complete assignment of the peaks besides the meta split peak for C7 (489 c/s, J7_5&2 c/s). The doublet without meta splitting at 502 c/s fits the C4-proton, and the C5-hydrogen doublet at 516 c/s showed the expected fine structure (J5_4 ^ 8 c/s). Since all the immunochemically significant fluoresceins (amines, isocyanates and isothiocyanates) are derived from the nitrofluoresceins by straightforward reactions, the positions of substitution are now known: the I series is 5-substituted, and isomer II compounds are substituted at C6. It is noteworthy that infra-red data led Borek2 to suggest structural assignments opposite to those which nuclear magnetic resonance so clearly imposes.We thank Dr. Roger M. McKinney, who supplied the nitrofluorescein derivatives.