We report Very Large Array radio observations of 29 supernovae (SNe) with ages ranging from 10 days to about 90 yr past explosion. These observations significantly contribute to the existing data pool on such objects. Included are detections of known radio SNe 1950B, 1957D, 1970G, and 1983N, the suspected radio SN 1923A, and the possible radio SN 1961V. None of the remaining 23 observations resulted in detections, providing further evidence to support the observed trend that most SNe are not detectable radio emitters. To investigate the apparent lack of radio emission from the SNe reported here, we have followed standard practice and used Chevalier's "standard model" to derive (upper limits to) the mass-loss rates for the supernova progenitors. These upper limits to the fluxes are consistent with a lack of circumstellar material needed to provide detectable radio emission for SNe at these ages and distances. Comparison of the radio luminosities of these supernovae as a function of age past explosion to other well-observed radio SNe indicates that the Type II SNe upper limits are more consistent with the extrapolated light curves of SN 1980K than of SN 1979C, suggesting that SN 1980K may be a more typical radio emitter than SN 1979C. For completeness, we have included an appendix where the results of analyses of the non-SN radio sources are presented. Where possible, we make (tentative) identifications of these sources using various methods.

Based on the results of VLA observations, we report the detection of two unresolved radio sources that are coincident with the reported optical position of SN 1923A in M83. For the source closest to the SN position, the flux density was determined to be 0.30 +/- 0.05 mJy at 20 cm and 0.093 +/- 0.028 mJy at 6 cm. The flux density of the second nearby source was determined to be 0.29 +/- 0.05 mJy at 20 cm and 0.13 +/- 0.028 mJy at 6 cm. Both sources are nonthermal, with spectral indices of alpha = -1.0 +/- 0.30 and -0.69 +/- 0.24, respectively. SN 1923A has been designated a Type II-P. No Type II-P (other than SN 1987A) has been detected previously in the radio. The radio emission from both sources appears to be fading with time. At an age of approximately 68 yr when we observed it, this would be the oldest radio SN (of known age) yet detected.

We have used the VLA to place deep upper limits on the radio emission from the Type II-L supernovae SN 1984E in NGC 3169 and SN 1986E in NGC 4302. The optical spectra of SN 1984E near maximum light showed evidence of strong circumstellar interaction. The lack of detectable radio emission from SN 1984E at an age of 11 years is consistent with the idea that near its optical maximum SN 1984E was interacting with a circumstellar shell that resulted from a preexplosion mass-loss episode that was of limited duration. However, optical spectra of SN 1986E have shown that it was undergoing strong circumstellar interaction as recently as 1994. Therefore, the lack of detectable radio emission from SN 1986E in 1995 is surprising. SN 1986E is the first "old" supernova to have been detected in the optical but not in the radio. This calls for further detailed modeling of the optical and radio emission from circumstellar interaction.

Searches for prompt radio emission from Type Ia supernovae (SNe Ia) have been proposed as a test of the symbiotic-star progenitor scenario, in which a white dwarf accretes mass from the wind of a companion red giant (Boffi & Branch 1995a). We report the results of a VLA search for radio emission from the Type Ia supernova 1986G in NGC 5128 (Centaurus A), 1 week before its optical maximum. These observations allow us to establish earlier and lower limits on SN Ia radio emission than previously available. Our 3 sigma upper limits of 1.0 mJy at 6 cm and 0.7 mJy at 2 cm indicate that the progenitor of SN 1986G probably was not a symbiotic star. SN 1986G was a peculiar Type Ia supernova, and further searches for prompt radio emission from other SNe Ia are needed to more thoroughly test the symbiotic-star progenitor scenario.