Observations across the electromagnetic spectrum of radiative processes involving interstellar dust—emission, absorption, and scattering—are used to constrain the parameters of dust models and more directly to aid in foreground removal of dust for extragalactic and cosmological observations. Dust models can benefit from more independent constraints from complementary observations. Here, we quantify the relationship between scattered light and thermal emission from dust in a diffuse (cirrus) intermediate-latitude cloud, Spider, using data from the Dragonfly Telephoto Array and the Herschel Space Observatory. A challenge for optical observations of faint diffuse cirrus is accurate removal of a contaminating, spatially varying sky. We present a technique to analyze two images of the same cirrus field concurrently, correlating pixel values to capture the relationship and simultaneously fitting the sky-related signal as a complex noncorrelating additive component. For the Spider, we measure a color g − r = 0.644 ± 0.024 and ratios of visible-wavelength to 250 μm intensity of γ
g,250 = (0.855 ± 0.025) × 10−3 and γ
r,250 = (1.55 ± 0.08) × 10−3 for the g and r-bands, respectively. We show how to use any dust model that matches the thermal dust emission to predict an upper limit to the amount of scattered light. The actual brightness of the cirrus will be fainter than this limit because of anisotropic scattering by the dust combined with anisotropy of the incident interstellar radiation field (ISRF). Using models of dust and the ISRF in the literature, we illustrate that the predicted brightness is indeed lower, though not as faint as the observations indicate.