## Abstract

H$_{3}$C$\chembond{1,0} $C\chembond{2,0} C$\chembond{1,0} $C$\equiv $C$\chembond{1,0} $C$\equiv $C$\chembond{1,0} $CH$_{3}$ crystallizes with one molecule in a rhombohedral cell of dimensions a = 5$\cdot $9 angstrom, $\alpha $ = 70 degrees 58$^{\prime}$. The space group is R$\overline{3}$m, and with the centrosymmetrical molecule extended along the triad axis four parameters define the positions of the carbon atoms. Because of the simplicity of the crystal structure, the X-ray analysis was refined to a degree of detail and accuracy not generally possible with organic molecular structures. The Fourier methods of refinement were applied independently to the (0, z) projection with (h0$\overline{h}$l) terms and the (0, 0, z) line with (hkil) terms. A sectional synthesis at (0, y, z) with (F$_{\text{obs}}$. - F$_{\text{calc.}}$) terms revealed no marked anisotropy of thermal vibration in the carbon atoms and showed a clearly defined peak of 0$\cdot $35 e angstrom $^{-3}$ corresponding to the location of the hydrogens on a circle coaxial with the molecule. A section at (x, y, z$_{\text{H}}$) revealed no evidence of other than a freely rotating methyl group. The agreement between observed and calculated structure factors was significantly improved by an allowance for the distribution of the $\pi $-bonding electrons. The Fourier syntheses were corrected for termination-of-series errors, after which the heights of the carbon peaks were identical within 0$\cdot $06 eA$^{-3}$, and the parameters from the two- and three-dimensional data were consistent within 0$\cdot $006 angstrom. The standard deviations of the carbon atomic co-ordinates calculated by the method of Cruickshank were 0$\cdot $010 and 0$\cdot $005 angstrom for projection and line respectively. The molecular dimensions from the three-dimensional analysis are: $\chembond{1,0} $C$\equiv $C$\chembond{1,0} $1$\cdot $19$_{9}$ angstrom, C$\chembond{1,0} $C 1$\cdot $37$_{5}$ angstrom, $\chembond{1,0} $C$\equiv $C$\chembond{1,0} $1$\cdot $19$_{9}$ angstrom, C$\chembond{1,0} $CH$_{3}$ 1$\cdot $46$_{6}$ angstrom, C$\chembond{1,0} $H $\sim $ 1$\cdot $0 angstrom, C$\chembond{1,0} $\hat{{\rm C}}$\chembond{1,0} $H 105 degrees. These data agree within experimental error with the spectroscopic measurements for the corresponding bond lengths in acetylene and methyl acetylene and the electron diffraction results for diacetylene and dimethyl acetylene. Assuming the validity of the $\pi $-bond order-length relationship of Lennard-Jones and Coulson in the region of the triple bond, there is a significant discrepancy between theory and experiment for the length of the central link.