The earliest systems of classification of ``nebulae'' by W. Herschel and others have been described by H. D. Curtis in his article of the ``Handbuch der Astrophysik'' [A] ¹. The spiral form of some nebulae was first detected visually by Lord Rosse and his assistants between 1845 and 1850 ² and the appellation elliptical first used in a purely descriptive way was applied to non-spiral tub presumably external stellar systems by S. Alexander in 1852 ³.

The great abundance of the spiral type was brought to light by the photographic surveys of the end of the XIX-th century. The spiral structure of the Andromeda nebula was detected photographically in 1888 by I. Roberts. These surveys also revealed the great abundance of smaller objects of the smooth, structureless elliptical type and the very elongated objects described as ``spindles''. In 1918 Curtis [9] isolated a special type of spiral, the so-called ``'' type, characterized by a diametral bar across the nucleus and a general ring-like structure; he also identified the ``spindle'' nebulae with a longitudinal dark lane as edgewise spirals with peripheral dark matter [10].

An early alphabetical classification of apparent shapes, introduced by M. Wolf ⁴, has been extensively used at Heidelberg by Reinmuth [62] and at Lund by Lundmark [44], Holmberg, Reiz, Danver and others ⁵. Its relation to the standard classification has been given by Shapley and Ames [68]. For faint nebulae which show little or no structure, a descriptive classification, based on ``concentration'' and ``elongation'', introduced by Shapley ⁶ in 1929, has been in use at Harvard for some years; it bears little relation to the standard classification.

Rather similar systems of morphological classification were introduced, about 30 years ago, by Hubble [25] and by Lundmark [43] with three main types, viz. elliptical, spiral (normal and barred) and irregular. This scheme, extensively used and later further developed by Hubble, has been accepted as standard up to the present time.