1: \begin{abstract}

2: We detect weak gravitational lensing of the cosmic microwave background (CMB) at the location of the WISE $\times$ SuperCOSMOS (\WxS) galaxies using the publicly available {\it Planck} lensing convergence map. By stacking the lensing convergence map at the position of 12.4 million galaxies in the redshift range $0.1\le z \le 0.345$, we find the average mass of the galaxies to be \mvir = \stackmasswithtwohaloterm\ $\pm$ \stackmassnotwohalotermerror\ $\times\ \mbox{10}^{12}\ \mbox{M}_{\odot}$. The null hypothesis of no-lensing is rejected at a significance of $\stacksigmawithtwohaloterm \sigma$. We split the galaxy sample into three redshift slices each containing $\sim$4.1 million objects and obtain lensing masses in each slice of \binonemasswithtwohaloterm\ $\pm$ \binonemassnotwohalotermerror, \bintwomasswithtwohaloterm\ $\pm$ \bintwomassnotwohalotermerror, and \binthreemasswithtwohaloterm\ $\pm$ \binthreemassnotwohalotermerror\ $\times 10^{12}\ \mbox{M}_{\odot}$. %Our results suggest a redshift evolution of the galaxy sample masses but this apparent increase might also be due to Malmquist bias in the galaxy catalogue. 

3: Our results suggest a redshift evolution of the galaxy sample masses but this apparent increase might be due to the preferential selection of intrinsically luminous sources at high redshifts. 

4: The recovered mass of the stacked sample is reduced by 28\% when we remove the galaxies in the vicinity of galaxy clusters with mass \mvir = $2 \times 10^{14}\ \mbox{M}_{\odot}$.

5: We forecast that upcoming CMB surveys can achieve 5\% galaxy mass constraints over sets of 12.4 million galaxies with \mbox{\mvir = $1 \times 10^{12}\ \mbox{M}_{\odot}$} at $z=1$.

6: 

7: \end{abstract}

8: