Note that in "propagation of errors" note number 2:

• (page 1) I have used the Einstein summation notation when I do the Taylor series expansion
• (page 2) D matrix is the matrix with elements <du_i du_j>

I noticed that the 1/pz term for t vs E in 2014-12-16_13.35.00.jpg should be 1/(pz beta^2 gamma ^2). 1/pz is the absolute value, not the derivative.

I note I dropped a factor 1/pz in the bottom left hand square of M_d in http://micewww.pp.rl.ac.uk/attachments/3173/2014-12-16_13.35.00.jpg

So elements m_31, m_32, m_41, m_42 (indexing from 1 through 4).