1: \begin{abstract}

2:   We develop a method for performing a weak lensing analysis using

3:   only measurements of galaxy position angles. By analysing the

4:   statistical properties of the galaxy orientations given a known

5:   intrinsic ellipticity distribution, we show that it is possible to

6:   obtain estimates of the shear by minimizing a $\chi^2$

7:   statistic. The method is demonstrated using simulations where the

8:   components of the intrinsic ellipticity are taken to be Gaussian

9:   distributed. Uncertainties in the position angle measurements

10:   introduce a bias into the shear estimates which can be reduced to

11:   negligible levels by introducing a correction term into the

12:   formalism. We generalize our approach by developing an algorithm to

13:   obtain direct shear estimators given any azimuthally symmetric

14:   intrinsic ellipticity distribution. We introduce a method of 

15:   measuring the position angles of the galaxies from noisy 

16:   pixelized images, and propose a method to correct for biases 

17:   which arise due to pixelization and correlations between 

18:   measurement errors and galaxy ellipticities. We also develop 

19:   a method to constrain the sample of galaxies used to obtain an estimate 

20:   of the intrinsic ellipticity distribution such that fractional biases 

21:   in the resulting shear estimates are below a given threshold value. 

22:   We demonstrate the angle only method by applying it to simulations 

23:   where the ellipticities are taken to follow a log-normal 

24:   distribution. We compare the performance of the position angle 

25:   only method with the standard method based on full ellipticity 

26:   measurements by reconstructing lensing convergence maps from 

27:   both numerical simulations and from the CFHTLenS data. We find that the 

28:   difference between the convergence maps reconstructed using the two methods 

29:   is consistent with noise. 

30: \end{abstract}

31: