Additional Wavelength Coverage

In this Section, we describe the details of how we obtained the photometric measurements at various wavelengths to determine the properties of the candidates from PaperI. The optical through far-IR photometry are reported in Table2, and the extended SEDs are shown in Figures 7 and 8.

We utilized VizieR (Ochsenbein et al.2000) to search for other observations of the candidates, in particular for WISE (Wright et al. 2010, $12\micron$), 2MASS (Cutri et al. 2003, $JHK_s$), SDSS (Abazajian et al. 2009, $ugriz$) and X-ray detections. For M33, we used the $UBVRI$ images from the Massey et al. (2006) optical survey, and archival HST images of NGC300, NGC2403, M81, NGC247 and NGC7793. Finally, we used Herschel PACS data to supplement the Spitzer measurements.

For the Spitzer IRAC 3.6, 4.5, 5.8 and $8\micron$ as well as MIPS (Rieke et al.2004) 24, 70, and $160\micron$ data, we use the measurements reported in PaperI. For M33, our measurements were based on IRAC data from McQuinn et al. (2007) and MIPS data from the Spitzer Heritage Archive. Data from the LVL survey (Dale et al.2009) were used for NGC300 and NGC247, and data from the SINGS survey (Kennicutt et al.2003) for NGC6822, NGC2403, and M81.

We used the Herschel PACS (Poglitsch et al.2010) 70, 100, and $160\micron$ images available from the public Herschel Science Archive. Although both MIPS and PACS cover the same far-IR wavelength range ($70-160\micron$), Herschel has significantly higher resolution (see Figure3). All three PACS band data were available for M33 and NGC7793, 70 and 160$\micron$ data were available for NGC2403 and M81, and 100 and 160$\micron$ data were available for NGC300. There are no publicly available PACS images of the candidates in NGC247. We used aperture photometry (IRAF ApPhot/Phot) with the extraction apertures and aperture corrections from Balog et al. (2013) and given in Table1. As with our treatment of the MIPS 70 and 160$\micron$ measurements in PaperI, we treat the measurements obtained in the PACS bands as upper limits because the spatial resolution of these bands requires increasingly large apertures at longer wavelengths. For similar reasons, we also treat the WISE $12\micron$ fluxes, where available, as upper limits.

For the optical photometry of the candidates in M33, we used the Local Group Galaxies Survey $UBVRI$ images (Massey et al.2006). First we verified that the coordinates match with the IRAC images to within few $\times0\farcs1$ and then used $1\farcs0$ radius extraction apertures centered on the IRAC source locations. We transformed the aperture fluxes to Vega-calibrated magnitudes using zero point offsets determined from the difference between our aperture magnitudes and calibrated magnitudes for bright stars in the Massey et al. (2006) catalog of M33.

For the candidates in NGC300, M81, NGC2403, and NGC247, we searched the ACS Nearby Galaxy Survey (ANGST, Dalcanton et al.2009) $B$, $V$ and (where available) $I$ band point source catalogs derived using DOLPHOT (Dolphin2000). We verified that the IRAC and HST astrometry of the NGC300, NGC2403 and NGC247 images agree within (mostly) $\lesssim0\farcs1$ to (in a few cases) $0\farcs3$. We corrected the astrometry of the M81 HST images using the LBT images described later in this section to achieve similar astrometric accuracy. We also used the HST $I$-band photometry of M81 from HST program GO-10250 (P.I. J.Huchra). We retrieved all publicly available archival HST images of NGC7793 overlapping the IRAC source locations along with the associated photometry tables from the Hubble Legacy Archive. The HST and Spitzer images have a significant (few $\times 1\farcs0$) astrometric mis-match, and there are too few reference stars in the HST images to adequately improve the astrometry. Therefore, we utilized the IRAF GEOXYMAP and GEOXYTRAN tasks to locally match the overlapping HST and Spitzer images of NGC7793 within uncertainties of $0\farcs1 \sim 0\farcs3$.

We have variability data for the galaxies M81 and NGC2403 from a Large Binocular Telescope survey in the $U B V R$ bands that is searching for failed supernovae (Kochanek et al.2008), and studying supernova progenitors and impostors (Szczygie et al.2012), and Cepheid variables (Gerke et al.2011). We analyzed 27 epochs of data for M81 and 28 epochs of data for NGC2403, spanning a 5year period. The images were analyzed with the ISIS image subtraction package (Alard2000; Alard & Lupton1998) to produce light curves (see Figure5).