Results from the photometric data reduction are summarized in Tables 2a and 2b (for the JKT and the ESO runs respectively). In the first column the name of the galaxy is given; then the supernova(e) and the SN Type are listed in columns two and three. In the fourth column we list all patches that have been identified at (or close to) the site of the supernova and that were bright enough to be detected during the observation: they are numbered and briefly described by a comment that follows in column five. Such features are close to/within a ring of approximately 5 arcsec in radius centered at the nominal position of the supernova (see Figs. 1 through 36). A major uncertainty in this work is the actual location of the historical supernova. Re-measurement of a small number of plates containing images of the actual event (ongoing work) suggests typical uncertainties of the order of a few arcsec. Hence, for the present purposes, we chose to study a region of $\rm {10} ~arcsec$ diameter centered at the nominal location of the SN. For each patch the offset coordinates (in arcsec) with respect to the SN position have been calculated and are given in the next two columns. The estimated B, V and R magnitudes and the B-V and V-R colors are listed in Cols. 8 through 12 of these tables. They are all corrected for local Galactic extinction. As mentioned in Sect 3.3, the following three columns give respectively the B, V and R limiting magnitudes corresponding to a ${\rm 5{\sigma}}$ detection (but some magnitudes correspond to a ${\rm 3 {\sigma}}$ detection limit and are indicated with an asterisk in Tables 2a and 2b). Finally the last three columns give the estimated ${\rm 1 {\sigma}}$ uncertainties of the measured B, V and R magnitudes. These uncertainties lead to a typical uncertainty of 0.2 magnitude in the colors B-V and V-R (see color/color plots to follow).

The positions of a handful of SNe are accurately known from radio observations or astrometric measurements. Pennington et al. (1982) provide accurate astrometric positions of SNe 1923A and 1957D (in NGC 5236). For other two supernovae in the same galaxy, SNe 1950B and 1983N, accurate positions are derived from radio observations (Weiler et al. 1986). Given these absolute positions, the supernovae were located with respect to the SN 1957D position, which was identified by looking at the SN field observed by Long et al. (1992). The position of SN 1960L is accurately known from astrometry (Porter 1993). In the case of SN 1989B we have collected a wide variety of positional information. The position of SN 1961V in NGC 1058 well agrees with the radio position given in Cowan et al. (1988). Finally the position of SN 1885A in NGC 224 (M31) is in very good agreement with the coordinates of its supernova remnant as measured from Hubble Space Telescope Wide Field Planetary Camera-2 images by Fesen et al. (1998). Hence, in these cases, we are much more confident of the location of the original event in our images.

**Table 2:** a) ^a The SN Type was taken from Branch 1990 and/or from the Sternberg Catalog; ":'' after a Type indicates that according to the Sternberg Catalog there is uncertainty; both Types are classified spectroscopically, but the Type II are also classified according to the shape of the light curve (P=plateau; L=linear) as given by the Sternberg Catalog. This means that whenever a Type II is indicated as a Type IIP or IIL the Sternberg classification was used; for some objects only the Sternberg classification Type is found and is indicated in parenthesis; for SNe 1940E and 1963L the two sources disagree: both are given; ^b the offsets are in respect to the location of the supernova and are in units of arcsec; ^c "indef'' indicates that the IRAF program "phot'' did not calculate the magnitude in some band and thus the color could not be derived; the (-) indicates no images were taken in that band. ^d limiting magnitudes corresponding to a $5 {\sigma}$ detection limit; an asterisk next to the magnitude indicates that the magnitude corresponds to a $3 {\sigma}$ detection limit: see text for details; ^e standard deviations relative to the observed B, V and R magnitudes

**Table 2:** b) ^a The SN Type was taken from Branch 1992 and/or from the Sternberg Catalog; ":'' after a Type indicates that according to the Sternberg Catalog there is uncertainty; both Types are classified spectroscopically, but the Type II are also classified according to the shape of the light curve (P=plateau; L=linear) as given by the Sternberg Catalog. This means that whenever a Type II is indicated as a Type IIP or IIL the Sternberg classification was used; for some objects only the Sternberg classification Type is found and is indicated in parenthesis; for SNEe 1961I, 1964F and 1967H the two sources disagree: both are given; ^b the offsets are in respect to the location of the supernova and are in units of arcsec. ^c "indef'' indicates that the IRAF program "phot'' did not calculate the magnitude in some band and thus the color could not be derived; the (-) indicates no images were taken in that band; "sat.'' for SN 1968L in NGC 5236 indicates that the R band was saturated near the nucleus. ^d limiting magnitudes corresponding to a $5 {\sigma}$ detection limit; an asterisk next to the magnitude indicates that the magnitude corresponds to a $3 {\sigma}$ detection limit: see text for details; ^e standard deviations relative to the observed B, V and R magnitudes

4.2 Description of supernova environments:

In the following we describe the SN environments presented in Figs. 1 through 36. All sites are carefully analyzed and plausible candidates indicated. The 5 arcsec radius ring centered at the nominal position of the supernova is depicted at each site. Each figure can be considered as a finding chart corresponding to each supernova. We also include some comparison to the same environments as observed with the Wide Field Planetary Camera-2 (WFPC-2) in archival Hubble Space Telescope (HST) proposals when available.

NGC 23 - SN 1955C: patch 2, south-west of the SN, is red in B-V and is within imaging limits from the ground. It is probably just part of the body of the galaxy.

NGC 210 - SN 1954R: one patch is clearly visible at the northern edge of the ring; it is very blue in both colors. Classified as a candidate (see Table 3).

NGC 224 - SN 1885A: five small patches ( $\le 1$ arcsec) have been detected. They all are red and bright (brighter than 20th V magnitude). The galaxy light might contaminate the environment.

NGC 253 - SN 1940E: two bright patches; in both cases the B-V is very red and the V-R very blue. Patch 1 appears compact and isolated; patch 2 is extended and arc-like. More structure is visible outside of the 5 arcsec ring and is possibly just part of the arm of the galaxy.

NGC 488 - SN 1976G: one very red and bright extended patch (far off North to the assumed SN position). It looks like a small tail "departing'' from the galaxy nucleus.

UGC 2069 - SN 1961P: no patches are distinguishable within the region just around the SN; an isolated compact object (possibly a star?) is to the south-east of the SN nominal position. An arc of compact bright objects lies well to the northern edge of the SN sections.

UGC 2105 - SN 1938A: two patches are located, one to the south-west of the supernova (within the $\rm 5 ~arcsec$ ring) and the other well further out in the same direction. The latter looks very blue from visual inspection. The photometry indicates that the first patch is blue in color, and that the second one is very blue and faint in visual magnitude. They are both brighter than ${\rm 3 {\sigma}}$ detection limits.

NGC 1003 - SN 1937D: two patches are identified in the immediate surroundings of the SN location: they are both displaced in the x coordinate from the SN position and are compact. The hosting galaxy was oberved with WFPC-2 on the HST. The region around the SN is not very crowded and compact sources of emission are visible. Two groups of stars are observed South of the SN immediate environment and appear unrelated to the region of interest.

NGC 1058 - SNe 1961V, 1969L: an elongated patch is clearly visible at the location of SN 1961V; in V it appears broken up into different smaller patches: a couple of these are faint and extremely blue in B-V. The region of SN 1969L is empty. SN 1961V classified as a candidate (see Table 3).

NGC 1073 - SN 1962L: no patches are visible visible in the immediate surroundings of the SN. An extended patch instead is visible off the $\rm 5 ~arcsec$ ring to the south-east of the SN position.

UGC 2259 - SN 1963L: a blue and compact object is visible to the south-east of the SN position on the edge of the ring. Two patches slightly off the ring, to the north-west, appear blue and patchy and should be further investigated. Classified as a candidate (see Table 3).

NGC 1325 - SN 1975S: no patches are visible either within or outside the ring.

NGC 2276 - SNe 1962Q, 1968V, 1968W: no R band observations are available. The first two SNe lie in crowded regions. South-east of SN 1962Q a bright and compact patch of light is seen. At least four relatively bright patches are identified at around the location of SN 1968V. Further observations are needed (in the R band for example). At the SN 1968W position some blue emission brighter than a ${\rm 3 {\sigma}}$ magnitude is visible. The last supernova is close to the nucleus of the galaxy.

NGC 6946/field A+field B - SNe 1917A, 1939C, 1948B, 1968D, 1969P, 1980K: at the sites of SNe 1948B, 1968D and 1980K some patches are located: they all are bright and quite blue. SN 1980K is likely the most interesting of all SNe in this galaxy; in fact from all the data collected at radio wavelengths over a long period of time (Weiler et al. 1986, 1991) and from the fact that its optical counterpart has also been identified, it appears a good candidate to look at because of dense circumstellar and interstellar environments. At the sites of SNe 1917A and 1969P bright blue patches are observed. Star-like objects appear in the environment. In the case of SN 1939C some objects are visible as well. SNe 1917A, 1969P and 1980K are classified as candidates (see Table 3).

NGC 7177 - SNe 1960L, 1976E: at the location of SN 1960L some patches are seen. A blue one is seen at the location of the supernova, brighter than a ${\rm 5{\sigma}}$ detection. As already discussed the position of SN 1960L is accurately known from Porter (1993) and thus the identification of a blue patch at the location of the supernova is of great interest and calls for further investigation. At the site of SN 1976E a very bright and red compact object is observed. SN 1960L is classified as a candidate (see Table 3).

NGC 7331 - SN 1959D: patch 1 is within a ${\rm 3 {\sigma}}$ detection limit; patch 2 is brighter and blue, although appears compact on the image and might be only a star. Patch 3 is also blue and compact. These patches are also visible on the archival WFPC-2 image that we analyzed. With WFPC-2 it is clear that there are groups of stars present (possibly new forming stars) and/or HII regions belonging to a galaxy spiral arm.

NGC 2935 - SN 1975F: one extremely bright compact object not too far off the SN position; red in B-V. It looks similar to what is seen at the location of SN 1976E in NGC 7177.

NGC 3115 - SN 1935B: one red and compact object is visible to the south-east of the SN location and at the edge of the ring. Several other patches of the same kind are found all around the region.

NGC 3627 - SNe 1973R, 1989B: in the case of SN 1973R a bright blue elongated object is visible. Almost at the location of SN 1989B a roundish fuzzy and very blue (in both B-V and V-R) patch of light is present. The emission we see is within a ${\rm 3 {\sigma}}$ detection limit. Both SN environments are part of spiral arms and very crowded. On the basis of our selection criteria, both supernovae are classified as candidates (see Table 3).

NGC 4038 - SNe 1921A, 1974E: no patches are visible within 5 arcsec of the nominal positions of these two supernovae.

NGC 4254 - SNe 1967H, 1972Q, 1986I: several bright and blue features are seen at the locations of these supernovae. The SN sites are quite crowded though and the visible features might simply belong to the spiral arms of the galaxy.

NGC 4303 - SNe 1926A, 1961I, 1964F: at all these sites some features are present. From visual inspection and its blue color, the faint arc-shaped patch at the SN 1926A site looks promising. It is brighter than a ${\rm 3 {\sigma}}$ detection limit. The extended, bright, round patch at the site of SN 1961I is quite likely to be a HII region or young open cluster. SN 1926A is classified as a candidate (see Table 3). Both SNe 1926A and 1964F are visible in the WFPC-2 images of the host galaxy. We believe we can confirm what we noticed in our ground based observations.

NGC 4321 (M100) - SNe 1901B, 1914A, 1959E, 1979C: all environments show relatively bright and blue patches of light. The SN 1979C environment is possibly the most promising, as it is known from radio observations to be characterized by a dense environment (Weiler et al. 1981; Weiler et al. 1986). This SN has also been optically identified. Only in this one case (in this galaxy) is an emission patch clearly visible within 5 arcsec from the SN nominal position. Less compelling evidence has been gathered for the other supernovae. SN 1979C is classified as a candidate (see Table 3). The positions of SNe 1959E and 1979C are also identified on some archival WFPC-2 observations. Multiple sources are present in the immediate surroundings of the SNe. Many may well be star clusters, others more directly related to the SN event. For example we know the optical counterpart to SN 1979C has been identified among all other sources (Van Dyk, private communication).

NGC 4382 - SN 1960R: two very faint patches are detected within the immediate SN regions. All images appear to be very noisy. Patch 1 is within a ${\rm 3 {\sigma}}$ detection limit.

NGC 4424 - SN 1895A: from both runs (March and May 1995) one faint and compact patch is visible to the south-east of the SN location (at the edge of the $\rm 5 ~arcsec$ ring); it is very red in B-V.

NGC 4674 - SN 1907A: an elongated structure lying to the south-west of the SN is visible. Some other patch is also visible far off the SN nominal position. Both are red in color and possibly are just part of the main body of the galaxy.

NGC 4753 - SN 1965I, 1983G: at both locations some structure is found. The March 1995 run is very noisy and does not provide any information. Only at the location of SN 1965I (May run) is some faint structure visible to the south-west of the supernova.

IC 4237 - SN 1962H: a bright, blue, roundish patch is visible in all sections and is located (within 5 arcsec from the SN position) at the end of an arc-like structure possibly part of a spiral arm of the galaxy. Classified as a candidate (see Table 3).

NGC5236 (M83) - SNe 1923A, 1950B, 1957D, 1968L, 1983N: all patches at all locations appear to be very bright and blue (both in the B-V and in the V-R colors). At the site of SN 1983N two compact patches are visible. Note that the positions of these supernovae are known accurately (see Sect. 4.1). In all these cases the SN star progenitors are likely to have been massive (two SNe are confirmed Type II's and one is a Type Ib; SN 1957D is likely to be a Type II, as possibly inferred from the radio data, see Weiler et al. 1986). Around such progenitors dense circumstellar and possibly interstellar environments are expected, thus allowing for the formation of light echoes. The best bets may be SNe 1957D and 1983N. SNe 1957D and 1983N are classified as candidates (see Table 3).

NGC 5253 - SN 1895B, 1972E: no R band observations are available. Some structure is present close to the location of SN 1895B. Some patches are also visible near the location of SN 1972E. SN 1895B is also on the edge of some WFPC-2 images. As single images in different filters were found the image quality of the summed image is not optimal. We can see some structure in the surroundings of the supernova but this HST imaging does not contribute additional information on the presence of reflected light.

NGC 5668 - SNe 1952G, 1954B: both show some faint structures slightly north of the SN positions. The surrounding environments are crowded. Both supernovae lie not too far from the nucleus of the galaxy and belong to spiral arms. The May observations are slightly affected by saturation; the March ones have a poorer signal-to-noise ratio (the galaxy was observed well into twilight). For the March run no R band observations are available. WFPC-2 observations resolve the underlying spiral arm structure. The environments are not particularly crowded and at the location of SN 1952G the same single bright object observed from the ground is also observed with HST, i.e. it remains unresolved with WFPC-2.

NGC 5857 - SN 1950H, 1955M: a compact and blue patch is found at the location of SN 1950H. This SN lies close to the nuclear part of the galaxy.

NGC 5861 - SN 1971D: a patch is visible within the $\rm 5 ~arcsec$ ring, very close to the location of the SN. Classified as a candidate (see Table 3). WFPC-2 observations show emission close to the SN position. Yet the HST images resolve the spiral arm and thus many compact objects are indeed visible, making it difficult to discriminate easily among them.

NGC 6181 - SN 1926B: a blue and bright double-lobed object is visible at the site of this supernova. It lies within the $\rm 5 ~arcsec$ ring and occupies more than half of it. Classified as a candidate (see Table 3). This SN lies on the top left corner of the PC images retrieved from the archive. The SN environment is not very crowded, at the extreme end of the spiral arm. Some faint compact sources of emission are visible.

NGC 6384 - SN 1971L: a blue arc-shaped patch is present just close to the center of the section and is brighter than the ${\rm 3 {\sigma}}$ detection limit. HST images resolve the underlying spiral structure and several compact sources of emission are present within the uncertainty ring. Some structure is seen around the nominal position of the SN.

IC 4719 - SN 1934A: along the circumference of the $\rm 5 ~arcsec$ ring a series of faint and blue knots are visible; some are within a ${\rm 5{\sigma}}$ detection and some within a ${\rm 3 {\sigma}}$ .

IC 4798 - SN 1971R: two bright patches are visible, one, patch 1, possibly belongs to the nuclear part of the galaxy. The second one is to the south-west of the SN.

NGC 6835 - SN 1962J: a patch is visible within the ring, to the east of the supernova. It is within a ${\rm 3 {\sigma}}$ detection limit.

NGC 7177 - SNe 1960L, 1976E: at the location of SN 1960L some patches are seen as in the JKT observations. A blue one is seen at the location of the supernova and is within a ${\rm 3 {\sigma}}$ detection limit. SN 1960L is classified as a candidate (see Table 3).

4.3 Candidate light echoes

On the basis of what we see at various SN positions, we classify some optical emission patches as likely candidate light echoes and present them in Table 3. These candidates are close to the SN nominal position (typically within the 5 arcsec ring, but some lie as far as 6 arcsec away), appear blue in color and are compact. A bonus is granted when the SN position is accurately known (see Sect. 4.1). All data in Table 3 are derived from Tables 1a through 2b.

**Table 3:** All data in this table were derived from Tables 1a through 2b
$\begin{tabular} {lccccccc} &&&&&&&\\ \hline \hline &&&&&&&\\ Galaxy & SN & SN Ty... ...$-$0.02\\ &(ESO)& &1&0.51E&0.59N&0.50&$-$0.30\\ \hline &&&&&&& \\ \end{tabular}$

$\begin{figure} \includegraphics [width=6cm]{ds8769f37.ps} \end{figure}$

Figure:37 Color/color plot for all patches at/around the location of the historical supernovae for both observing runs. In this figure only patches whose B, V and R magnitudes are brighter than the calculated limiting magnitudes (corresponding either to a ${\rm 5{\sigma}}$ or a ${\rm 3 {\sigma}}$ detection, as presented in Sect. 4.1) are plotted. The ESO observations are the open circles ( ${\rm 5{\sigma}}$ ;40 objects) and the open squares ( ${\rm 3 {\sigma}}$ ; 11 objects); the JKT are the filled circles ( ${\rm 5{\sigma}}$ ; 37 objects) and the filled squares ( ${\rm 3 {\sigma}}$ ; 2 objects); SN 1991T is represented by the filled triangles (data from Schmidt et al. 1994)

$\begin{figure} \includegraphics [width=6cm]{ds8769f38.ps} \end{figure}$

Figure 38: The same as Fig. 37 but the corresponding B-V and V-R error bars are plotted. The corresponding standard deviations in the three band of observations are given in Tables 2a and 2b

$\begin{figure} \includegraphics [width=6cm]{ds8769f39.ps} \end{figure}$

Figure:39 Figure 39 is the same as Fig. 37 but allowing for a differentiation of the SN Types: all filled symbols refer to Type I SNe (Ia's), all open ones to Type II's (the only Type Ib SN fulfilling the selection criteria for this plot is SN 1983N and is plotted among the Type II as they likely derive from similarly massive star progenitors; in our samples there are no Type Ic SNe). The ESO observations are the squares; the JKT the circles. The empty triangles refer to peculiar supernovae and those whose Type is unknown according to the classification in Tables 2a and 2b. The filled triangles are SN 1991T (data from Schmidt et al. 1994)

$\begin{figure} \includegraphics [width=6cm]{ds8769f41.ps} \end{figure}$

Figure:41 The same as Fig. 39. The reddening vector (solid line), derived from the standard reddening law of Rieke & Lebofski (1985), and the scattering vector are also plotted. See text for details

4.4 Correlations and SN environments

Figure 39 is the same as Fig. 37 but allowing for a differentiation of the SN Types: all filled symbols refer to Type I SNe (Ia's), all open ones to Type II's (Type Ib's are inserted in this group as they likely derive from similarly massive star progenitors; in our samples there are no Type Ic SNe). The ESO observations are the squares; the JKT, the circles. The empty triangles refer to peculiar supernovae and those whose Type is unknown according to the classification in Tables 2a and 2b. The filled triangles are SN 1991T (from Schmidt et al. 1994). We have always followed the spectroscopic classification of Branch (1990) and therefore SNe 1961I, 1961V and 1964F have been considered Type II, while SNe 1940E, 1960L, 1963L and 1967H unknown. SN 1961V (in NGC 1058) which was classified as a Type V by Zwicky (1965) might not be a pure SN event (Goodrich et al. 1989; Filippenko et al. 1995) and yet can be as useful as any other object if it produced a light echo. It is clear that different Types of SNe are located in different regions of the color/color plot: the emission patches associated to Type Is are mostly located at B-V and V-R redder than $\sim{0.10}$ ,while Type IIs are on average bluer than this value in both colors. Unknown/peculiars are randomly distributed; the location of the peculiar Type Ia SN 1991T is somewhat consistent with the other peculiars. This is what might be expected from the distribution of different Type SNe in different galaxies and in different parts of the parent galaxies themselves. SNe II/Ib likely derive from massive progenitors in spiral galaxies and thus are associated to star-formation regions where a strong H $\alpha$ emission is present (Van Dyk et al. 1996 and references therein). This would tend to cause a high positive value of (V-R). Type Ia SNe, instead, possibly derive from a wider variety of stellar environments (as they are observed both in elliptical and spiral galaxies; see Barth et al. 1996 for references).

In Fig. 40 we plot all supernovae except the unknown/peculiar. We note that known Type SNe cover a much smaller portion of the color/color plane than unknown/peculiar.

In Fig. 41 we plot for comparison both a reddening vector (solid arrow) and a scattering vector (dotted arrow) to ascertain in what proportions these two physical components affect the colors of our sample of patches. The reddening vector was derived by assuming the Rieke & Lebofski (1985) reddening law: E(B-V)/E(V-R)=0.78. The effect of scattering, also, was estimated using scattering cross-sections as tabulated in Sparks (1994). With these numbers and assuming wavelengths of 4400, 5500, 6400 for the BVR filters, light which has been scattered is always bluer by $\rm {-0.3}$ in B-V and $\rm {-0.2}$ in V-R than the unscattered illuminating source. For illustration, the scattering vector shown originates at B-V=V-R=0, which is a reasonable estimate of the colors of a supernova at peak luminosity in many cases (D. Branch, private communication). In fact, we cannot be sure in detail what the colors of the time integrated supernova light curves are since decay rates can be quite wavelength dependent, and reddening may also play a role. Nevertheless, outliers, and especially blue outliers, in these diagrams should be considered as objects worthy of additional follow-up observations as potential echoes, and yet strict SN 1991T look-alikes, i.e. definite echoes (Sparks et al. 1999), are not present in the sample.

Finally, for reference, in Fig. 42 colors of Main Sequence stars of all spectral types were plotted over our sample (these colors were taken from Zombek 1992). The distribution of patches is broadly consistent with the locations of these Main Sequence stars. This distribution covers also a similar region to a composite stellar population as a function of age as was found by plotting an evolutionary track kindly provided by C. Leitherer. We therefore infer that for the most part we are simply seeing stars or star clusters. However there are in addition patches with both unusually red and unusually blue colors which should be considered with special care.

4 Results

4.1 Basic data

4.2 Description of supernova environments:

4.3 Candidate light echoes

4.4 Correlations and SN environments