An arc of calcium burning in hydrogen at low pressure emits numerous bands in the region 6000-7000A. This spectrum was photographed at high dispersion and two groups of bands, shading toward the violet may be distinguished: the A groups with heads at $\lambda \lambda 7035$, 7028, 6921, 6903 and the B groups with heads at $\lambda \lambda 6389$, 6382. In addition to these, the arc emits an isolated C group—a single band in the ultra-violet at $\lambda 3533.6$. This group is identical with a band of calcium hydride recently studied by R. S. Mulliken. The structures of A, B and C are very different. The A group forms a doublet system (${\mathrm{A}}_1$, ${\mathrm{A}}_2$) of $P-Q-R$ branches. The bands of the B group have a similar structure to that of the violet cyanogen bands, signified by doublet $P_1$, $P_2$ and $R_1$, $R_2$ branches. The C group consists of a single band having $P-R$ branches. In all bands the series deviate largely from polynomials of second degree. Thus, in B and C there is a remarkable "red-shift" of high numbered lines, accompanied by a sharp cut-off in their intensity. From combinations found between the $P-R$ branches, conclusions are reached regarding the spectral terms in CaH. The A, B and C groups have a common final ($N$) electronic term with a rotational doubling (${\epsilon }_2=\pm \frac{1}{2}$, ${\sigma }_2=0\mathrm{}$). The initial state of A forms an electronic doublet (${\mathrm{A}}_1$, ${\mathrm{A}}_2$) with the emission electron in a $\sigma$-orbit (${\epsilon }_1=0\mathrm{}$, ${\sigma }_1>0\mathrm{}$), thus explaining the appearance of $Q$ branches in A. In B (initial) there is again a rotational doubling (${\epsilon }_1=\pm \frac{1}{2}$, $\sigma =0\mathrm{}$). In C (initial) only one $\epsilon$ component is present (${\epsilon }_1=-\frac{1}{2}$, ${\sigma }_1=0\mathrm{}$). The departure from half-integral quantum numbers in C is avoided by accepting a large Kratzer's linear term $2\delta j$. The nuclear spacings in the CaH molecule are not in correlation with their vibration frequencies, violating a rule by Birge and Mecke. A comparison of the A group with the spectra of ZnH, CdH and HgH shows several interesting parallels, confirming the theory of Mulliken regarding these spectra.

• Received 30 September 1926

DOI:https://doi.org/10.1103/PhysRev.29.97