A distance Comparison principle for evolving curves (Huisken)

Question

I am reading "A distance Comparison principle for evolving curves", an article where Huisken gives an alternative proof of Grayson's theorem. I can't understand the proof of Theorem 2.1. Why $0=\delta(e_1\oplus 0)(d/l)=\frac{d}{l^2}-\frac{1}{l}\langle w,e_1\rangle$ and $0\leq\delta^2(e_1\oplus e_2)(d/l)=\frac{1}{l}\langle \omega, k(q,t_o)-k(p,t_0)\rangle$ ?