As you can see in the following source example, when set_xlim(0.0,5) is called, negatives values are still plotted in the polar plot. Whereas if set_xlim(0.01,5) or set_xlim(-0.01,5) are called, the drawing are as expected.

from matplotlib.pyplot import figure, show
import numpy as np

x = np.arange(0,1,0.001)
y = np.arange(-2.0,0.0, 0.002)
fig = figure()

ax = fig.add_subplot(131, polar=True)
ax.plot(x, y)
ax.set_ylim(bottom= 0.0, top=5)

ax = fig.add_subplot(132, polar=True)
ax.plot(x, y)
ax.set_ylim(bottom= 0.01, top=5)

ax = fig.add_subplot(133, polar=True)
ax.plot(x, y)
ax.set_ylim(bottom= -0.01, top=5)

fig.show()
raw_input()