The so-called Wurster's salts, radicals which occur in the oxidation of the paraphenylene diamines, show a strong absorption in the visible. The frequency and the intensity were calculated by the method of completely antisymmetrized products of molecular orbitals, with the introduction of but one empirical constant, namely, the difference between the ionization potential of the non-bonding nitrogen electron in ammonia, and the ionization potential of benzene. The results are in good agreement with the observed absorption band. The calculated intensity corresponds to an electron number of about 0.3, the observed one to about 0.1. The shift of the band to longer wave-length, and the intensification upon substitution of C${\mathrm{H}}_3$ groups on the nitrogen is explained qualitatively.

The infra-red spectra of ${\mathrm{C}}_2$${\mathrm{H}}_2$, HCN, and C${\mathrm{O}}_2$ give evidence of a splitting of the II vibrational levels (in which a ⊥ vibration is singly excited) into two components which is entirely analogous to the $\Lambda$-type doubling of the electronic states of diatomic molecules and is called here $l$-type doubling. The experimental data show that this $l$-type doubling may be represented by $\Delta \nu =q_{i}J(J+1\mathrm{})\mathrm{}$ and that the constant $q_i$ is of the same order as the constant ${\alpha }_i$ in $B_{\mathrm{}\left[v\right]\mathrm{}}=B_e-\Sigma {\alpha }_i(v_i+\frac{1}{2}{d}_i)$. The existence of the $l$-type doubling also explains the apparent discrepancy that the fundamental bands of HCN and ${\mathrm{C}}_2$${\mathrm{H}}_2$ do not show a convergence of the $P$ and $R$ branches while the $Q$ branch is distinctly shaded to shorter wave-lengths. A preliminary theoretical treatment of the $l$-type doubling is also given. The doubling is due mainly to the fact that in the displaced position of a perpendicular vibration a linear molecule is a slightly asymmetric top, but to some extent also to the Coriolis interaction of different vibrations. The observed values of $q_i$ are all larger than the calculated ones. Taking the $l$-type doubling into account the rotational constants ${\alpha }_i$ of C${\mathrm{O}}_2$ have been newly evaluated and from them and the value $B_{000}={0.3895}_0$ ${\mathrm{cm}}^{-1\mathrm{}}$ the rotational constant $B_e$ in the equilibrium position is found to be 0.3906±0.0002 ${\mathrm{cm}}^{-1\mathrm{}}$ which gives for the moment of inertia $I_e=71.61\mathrm{}·{10}^{-40\mathrm{}}$ g ${\mathrm{cm}}^2$ and for the CO distance $r_e=1.1615\mathrm{}·{10}^{-8\mathrm{}}$ cm.

Existing knowledge of the forms and the electronic spectra of a series of triatomic oxide and similar molecules BAB are briefly reviewed. The evidence (mostly from infra-red data) which establishes the acute-angled model of ozone as very probably correct is reviewed, one or two new points being added. New conclusions concerning the probable electronic structure of ozone are given. A new diagram is introduced showing the ionization energies of the various normal and excited valence-shell molecular orbitals as a function of the apex angle BAB. It is shown that this diagram, or a similar diagram, may be very useful in understanding the electronic structures and spectra of the molecules considered. The diagram makes possible a number of predictions which may be tested experimentally. The forms of the molecular orbitals of C${\mathrm{O}}_2$ are reviewed, with some interesting new modifications.

The structure of the $\lambda 4050$ group in comets appears to be incompatible with the assumption of a diatomic emitter. Rather, the structure is in conformity with that expected for a ⊥ band of a nearly symmetric top molecule if the moment of inertia about the top axis is approximately 0.35×${10}^{-40\mathrm{}}$ g ${\mathrm{cm}}^2$. Such a small value is possible only for a slightly bent X${\mathrm{H}}_2$ molecule with X = C, N, or O. For C${\mathrm{H}}_2$ and N${\mathrm{H}}_2^{+}$ a ⊥ band is to be expected in the region 4500-4000A. Of these two possibilities C${\mathrm{H}}_2$ is the most likely. Since the CH radicals observed in the comets must necessarily be formed from saturated hydrocarbons by successive photodecompositions one should indeed expect to find the spectra of intermediate molecules that lie in the accessible region.