Proof:

We shall denote the ends of $ \Delta_1$ by $ a$ and $ b$. If $ \sigma:\Delta_1\longrightarrow X$ is a singular one simplex then $ \partial _1\sigma = \sigma(b) - \sigma(a)$ which is obviously in ker$ \;\epsilon$ and we conclude that $ B_0(X)\subset$   ker$ \;\epsilon$. To prove the reverse inclusion, let $ \sigma$ be an arbitrary element of ker$ \;\epsilon$ given by (31.1). That is, the coefficients satisfy $ c_1 + c_2 + \dots + c_k = 0$. Pick any point $ p \in X$ and for each $ j$ let $ \sigma_j:\Delta_1\longrightarrow X$ be a path in $ X$ joining $ p$ and $ p_j$. We claim that $ \sigma$ is the boundary of the one chain $ \tau = c_1\sigma_1 + c_2\sigma_2 + \dots + c_k\sigma_k$.
$\displaystyle \partial _1\tau$ $\displaystyle =$ $\displaystyle c_1(\sigma_1(b) - \sigma_1(a)) + c_2(\sigma_2(b) - \sigma_2(a)) + \dots +
c_k(\sigma_k(b) - \sigma_k(a))$  
  $\displaystyle =$ $\displaystyle (c_1p_1 + c_2p_2 + \dots + c_kp_k) - (c_1 + c_2 + \dots + c_k)p = \sigma.$  

The last part follows from the fundamental theorem on group homomorphisms.

nisha 2012-03-20