Although deliberate effort was made to find a descriptive classification which should be entirely independent of theoretical considerations, the results are almost identical with the path of development derived by Jeans[8] from purely theoretical investigations. The agreement is very suggestive in view of the wide field covered by the data, and Jeans’s theory might have been used both to interpret the observations and to guide research. It should be borne in mind, however, that the basis of the classification is descriptive and entirely independent of any theory. _Elliptical nebulae._—These give images ranging from circular through flattening ellipses to a limiting lenticular figure in which the ratio of the axes is about 1 to 3 or 4. They show no evidence of resolution,[9] and the only claim to structure is that the luminosity fades smoothly from bright nuclei to indefinite edges. Diameters are functions of the nuclear brightness and the exposure times. [Illustration: PLATE XII Elliptical and Irregular Nebulae E0 NGC 3379, E2 NGC 221 (M 32), E5 NGC 4621 (M 59), E7 NGC 3115, NGC 3034 (M 82), NGC 4449] The only criterion available for further classification appears to be the degree of elongation. Elliptical nebulae have accordingly been designated by the symbol “E,” followed by a single figure, numerically equal to the ellipticity (a – b)/a with the decimal point omitted. The complete series is E0, E1, ..., E7, the last representing a definite limiting figure which marks the junction with the spirals. The frequency distribution of ellipticities shows more round or nearly round images than can be accounted for by the random orientation of disk-shaped objects alone. It is presumed, therefore, that the images represent nebulae ranging from globular to lenticular, oriented at random. No simple method has yet been established for differentiating the actual from the projected figure of an individual object, although refined investigation furnishes a criterion in the relation between nuclear brightness and maximum diameters. For the present, however, it must be realized that any list of nebulae having a given apparent ellipticity will include a number of tilted objects having greater actual ellipticities. The statistical average will be too low, except for E7, and the error will increase with decreasing ellipticity.