These two method calls should perform the same log scaling action. However, the output from xaxis.set_scale('log') behaves differently. See the attached figures for the difference. I am guessing yaxis counterparts behave the same way.

fig = plt.figure(figsize=(5, 5))
ax1 = fig.add_subplot(1,1,1)
ax1.plot(np.random.randn(100))

ax1.xaxis.set_scale('log')
ax1.set_xscale('log')

plt.matplotlib.__version__ 
'1.3.x'