I. Introduction. No consideration appears to have been given in the literature to the design of beam apparatus for studying the atomic spectra of gases. Considering our uniquely complete theoretical knowledge of the spectrum of hydrogen, and the inconclusive nature of our present experimental knowledge of the structure of the alpha-line, atomic beam studies of this line are needed. II. Geometrical consideration. The pressure ratio of the final and initial chambers must be far smaller than the solid angle subtended by the first slit at a point in the field of view; and the region of space in which the atoms are excited should not extend, in the line of view, far beyond the full (non-penumbral) portion of the beam. III. The Doppler effect. Five independent contributions to the Doppler effect, for a line emitted from an atomic beam, are studied: (1) Original atomic velocity; (2) atomic velocity gained by excitation (in particular, for hydrogen excited by electron impact); (3) atomic velocity gained by emission; (4) spread of light in the optical system; (5) error of orientation of the optical system with respect to the beam. Some of these factors tend to produce broadening, and some to produce frequency shifts. The broadening can probably be kept low enough for the complete resolution of the electronic structure of the alpha-line, $\frac{\nu }{\Delta \nu }=4.2\mathrm{}\times {10}^5$. IV. Design of the apparatus. An atomic beam apparatus is described which has been designed and built with a view to the conditions outlined above.

A critical survey is made of existing knowledge on the absorption spectra near $\lambda 2000$ of ethylene, butadiene, and their alkyl derivatives. Particular emphasis is given to the nature of the molecular structures and electronic states involved. The results of some preliminary quantum-theoretical computations of C. A. Rieke and the writer on electronic term values and on red shifts due to hyperconjugation with alkyl groups are reported. It is concluded that decreases in ionization potential on alkyl substitution are due mainly to charge transfer as suggested by Price, but that red shifts in spectra result partly from this and partly from hyperconjugation. Various points raised in papers of Price and collaborators, and in the paper of Carr, Pickett, and Stücklen (CPS) in this issue, are discussed. The spectroscopic and refractivity data, combined with the computations, favor the $s$-trans form as the stable form in butadiene and its open-chain derivatives, but there is some evidence pointing to the presence of admixture of the $s$-cis form; further study is needed. (The nomenclature $s$-cis and $s$-trans is here proposed for the arrangement of groups around a single bond.) The stability of the planar arrangement in various electronic states of the molecule and its positive ion is discussed; the evidence indicates that the planar form is barely stable in the ion. Attention is called to the need of further study of the phenomenon of photo-chemical cis↔trans conversion by light near $\lambda 2000$.