Annu. Rev. Astron. Astrophys. 1997. 35: 309-355
Copyright © 1997 by . All rights reserved

Next Contents Previous


This review has emphasized the optical spectral classification of SNe. The approach is largely taxonomical - or, in the style of Zwicky (1965), "morphological." Like botanists and zoologists, we wish to find observable characteristics that eventually provide a deeper physical understanding of the objects under consideration. Naturally, certain properties may turn out to be more useful than others, and it is our goal to identify these.

For example, a clue to whether SNe Ib and SNe Ic are physically different might be provided by the degree to which the He I line strengths form a continuous sequence among the two subclasses; a roughly bimodal distribution could indicate distinct progenitors or evolutionary paths. Similarly, if the observed properties of SNe IIn really do result from unusually dense circumstellar media, they will provide information on stellar mass loss under different circumstances. The apparent similarity of the early-time spectra of SNe Ic and subluminous SNe Ia (Filippenko et al 1992a), on the other hand, may end up being nothing more than an indication of comparable primordial iron abundances in the atmospheres of the progenitors, as suggested by Clocchiatti & Wheeler (1997).

There is a tendency to assign each object to a new pigeonhole, based on small variations in the spectra or light curves. This should generally be resisted, unless there are clear physical grounds for doing so (as in the case of SNe IIb), because the proliferation of subtypes having few known members generally does not enhance our understanding of the nature of the phenomenon. For instance, the unusual strength of Ba II lines in early-time spectra of SN 1987A (Williams 1987), though interesting and important to understand, does not justify the creation of a new subclass. Of course, deviants often give valuable clues that are otherwise difficult to notice or entirely unavailable, and considerable attention should be paid to them. Good examples are provided by the peculiar SNe Ia 1991T and 1991bg, which dramatically illustrate the heterogeneity of SNe Ia and considerably strengthen the apparent correlation between luminosity and decline rate.

The greatest benefits, of course, are often achieved when many different types of observations are combined. For example, optical polarimetry and spectropolarimetry, though not discussed here owing to space limitations, can give information on asymmetries in the ejecta of SNe (e.g. Jeffery 1991, Höflich et al 1996b). Similarly, the recent work of Clocchiatti & Wheeler (1997) showed that SNe Ic with essentially identical spectra can be distinguished by their optical light curves, perhaps indicating rather important differences in the internal structure of the ejecta. IR emission is prominent throughout the evolution of SNe and thoroughly dominates at late times. Conditions at the time of shock breakout are best studied at UV wavelengths, X-ray and radio observations provide clues to the circumstellar environment of SNe, gamma-ray data are used to study the products of explosive nucleosynthesis, and neutrinos can indicate whether a neutron star was formed. The observational study of SNe is a rich and active field; this review was necessarily restricted to only a small subset of the relevant data.


Many colleagues have contributed to my knowledge of the spectra of SNe, sharing their insights in spirited conversations and stimulating lectures, as well as in comments on a draft of this review. Some of my students and former students (AJ Barth, LC Ho, DC Leonard, T Matheson, and JC Shields) helped obtain and calibrate the Lick spectra shown here; special thanks go to DC Leonard for his assistance with many of the figures. S Benetti and A Clocchiatti kindly sent their data in digital format. My research on SNe has been supported by the National Science Foundation, most recently through grant AST-9417213.

Next Contents Previous