Compare also our hints to
recent developments,the lists of
misprints and our
list of recent literature.
The following text contains characters from the
"symbol" font, to display mathematical formulae.
It may happen that these symbols appear crambled
on your screen, or that your browser is able
to show them right on the screen, but is unable
to print them correctly.
It may also happen that the characters on
your screen appear to be perfectly innocent,
but their alignment is not quite right.
In these cases
(and in order to produce a nice printout),
we offer the text also as a
PostScript version
and as a DVI version.
Consult
this link to find information about enabling
the symbol fonts with netscape under X.
Click here to go directly to remarks on Chapter 1, Chapter 2, Chapter 3, Chapter 4, Chapter 5, Chapter 6, Chapter 7, Chapter 8, Chapter 9, or the Bibliography.
32.2: It should be noted that the `pencil intersection' topology PI has to be contained in any topology that turns the projective completion into a topological plane. As we are dealing with compact Hausdorff spaces, this implies that PI is the only possible topology for the point space. The onepoint compactification of the line space is also the only reasonable topology for the line space of the projective completion.
This is needed in 32.5 and 42.10 to make sure that the given topology on P coincides with the topology constructed from the affine plane.
32.5: The proof needs an additional argument to show that the topology on P is the `pencil intersection' topology PI with respect to the affine plane on P\L; see the remarks on 32.2.
34.4: In part 2) on page 189/190, it is impossible to describe both oneparameter groups E and X by the same formula. Instead, the definitions should be



42.10: The proof needs an additional argument to show that the topology on P is the `pencil intersection' topology PI with respect to the affine plane on P\L; see the remarks on 32.2.
43.8: The hypothesis that the map t ® f(s,x,t) of R^{n} be injective should be made for every s, x Î R^{n} (and not only for s = 0). This is needed in the last paragraph but one of the proof. The error does not affect the applications of (43.8) made within the book.
There is another gap in the proof, but that one can be remedied. The third paragraph tacitly uses that lines defined by t = 0 pass through the origin (0,0). This is not implied by the assumptions. It can be enforced, however, by preceding the proof given with one further step of transformation. One wants to achieve that f(s,0,t) = t for all s, and one uses the inverse of the homeomorphism (s,t) ® (s,f(s,0,t)). Let this inverse be given by (s,t¢) ® (s,y(s,t¢)); then one has to replace f by
