Errata For Enumerative Combinatorics, Vol. 1, 2nd Ed., 2012
Errata For Enumerative Combinatorics, Vol. 1, 2nd Ed., 2012
• page 34, line 15. It should be stated that for this second way, d + 1
cannot be inserted at the end even though d 6∈ D(w).
• page 40, line 15. The intent of the parenthetical comment was to write
the summation as
X X X
··· q g(4)+···+g(5) .
g(4)≥g(6) g(6)≥g(3) g(5)≥0
• page 49, lines 20–22. The two sentence beginning “Figure 1.10 . . . ”
and “Let f (n) denote . . . ” should be interchanged, since f (n) is used
in the first of these sentences but defined in the second.
• page 98, line 7. Change “in the monograph” to “is the monograph”.
• page 109, Exercise 35(b,c,d). For the history of these results, see Math-
Overflow 63561.
• page 112, Exercise 54(a), line 1. Change A(2n+1, n+1) to A(2n+1, n).
• page 115, Exercise 70(b), line 3. Change (1, 0) to (0, 1), and change
“i.e,” to “i.e.,”.
• page 140, Exercise 199, line 4. The condition (only relevant in charac-
teristic 2) should be that A = −At and A has 0’s on the main diagonal.
Some people will take this to be the definition of “skew-symmetric,”
while others will call these matrices “alternating.”
• page 163, Exercise 80. For the convoluted history of this result, see R.
Gilbert, A Fine rediscovery, Amer. Math. Monthly 122 (2015), 322–
331. In particular, the result is actually first due to N. J. Fine in 1959.
• page 166, Exercise 92. This identity is actually due to I. Gessel and
D. Stanton, Trans. Amer. Math. Soc. 277 (1983), 173–201 (equations
(7.13) and (7.15)).
• page 169, Exercise 109(a), line 4. Change “is bijection” to “is a bijec-
tion”.
• page 171, line 4. The index entries for these four names is missing.
• page 173, Exercise 134, line 2. Change the period after Petersen to a
comma.
• page 210, line 9. Insert “one for 121,” after “one for 112,”.
• page
P 227, line 3–. iOn the left-hand side of the formula, remove one
i
i≥0 and change y to t .
• page 253, line 9–. In the definition of finitary distributive lattice, add
the condition that L has finitely many elements of each rank. (Thus
for instance the distributive lattice of all finite subsets of an infinite set
is locally finite with 0̂ but not finitary.)
• page 277, line 12–. Change dim(W ∪ W ′ ) to dim(W + W ′ ).
• page 284, line 3–. Delete “real” at the end of the line.
• page 285, line 11. Under the second Σ, change B1 ∈ A′′ to B1 ⊆ A′′
• page 315, line 7. Change as − as+1 to as − as−1 (twice after the brace).
• page 317, after (3.80). The phrase “(since intervals of Eulerian posets
are Eulerian)” is unnecessary since the formulas under consideration
hold for any graded poset with 0̂ and 1̂.
• page 337, lines 14– to 11–. The sentence “For the algebraically minded
. . . formal power series.” is not correct. We first need to consider non-
commutative polynomials (not power series) and then pass to suitable
completions.
• page 342, line 6. Change (1 − q)2 to (1 − q) (in the denominator).
• page 345, line 4. Change ps−1 + ps to pj−1 + pj .
• page 345, line 13. Change A to A (boldface).
• p. 356, Exercise 15(g), line 3. Change 9655 to 11586.
• page 356, Exercise 16(b). Change the difficulty rating to [3] (or maybe
[3+]). See M. Guay-Paquet, A. H. Morales, and E. Rowland, DMTCS
proc. AS (FPSAC 2013 Paris) (2013), 253–264.
• page 360, Exercise 38. This exercise is incorrect for 1 < k < m, where
m is the maximum size of an antichain of P . A counterexample is given
by k = 2 and L = J(P ), where P = 1 + (1 ⊕ (1 + 1)). It is true that
#Pk = #Rk , even for modular lattices. See Exercise 3.101(d).
• page 360, Exercise 41(c,d). It is assumed that L is distributive.
• page 364, Exercise 55(b). The rating should be changed to [3–]. An ex-
ceptionally elegant proof was given by G. Stachowiak, Order 5 (1988),
257–259. Another elegant proof was given by B. Iriarte, arXiv:1405.4880.
Iriarte proves the more general result that if G is the comparability
graph of a poset P , then the number of linear extensions of (G, o) is
maximized when o respects the order P .
• page 374, Exercise 95. The stated identity is actually valid for the dual
Möbius algebra, where one uses the join operation rather than the meet.
The correct identity using the text definition of Möbius algebra is
!
X X X
µ(t, 1̂)t = µ(u, 1̂)u · µ(v, 1̂)v .
t∈L u≥z v∨z=1̂
(x − 1)(x − 2) · · · (x − n).
n
• page 379, Exercise 115(c), line 1. Change “0,1” to 0. (There are 2
hyperplanes.)
• page 387, Exercise 3.139. The right-hand side of the displayed formula
should be 2x(1 + x)n−2 /(1 − x)n .
• page 390, Exercise 158(c). Delete one of the periods at the end of the
line.
• page 403, Exercise 189. Note. The result is also true for d odd, but
the proof is quite a bit easier.
• page 405, line 8–. The notation f00 is not defined until the next sen-
tence.
• page 407, Exercise 205(b), line 2. Change 44605 to 44606 (private com-
munication from Patrick Byrnes, 21 February 2012). Byrnes originally
assumed that a vertex v could be covered by at most one singleton
(element covering only v), but there is exactly one example up to rank
9 where this property need not hold. Byrnes also computes that there
are 29,199,636 1-differential posets up to rank 10.
• page 408, Exercise 215(c). Change the difficulty rating [5] to [3].
• page 415, Exercise 36(b). This example is incorrect. For instance,
{1, 2, 3} and {1, 2, 4, 5} have no meet.
• page 522, Figure 4.29. The vertex labels should be 0 and 1, not 00 and
10.
• page 530, Exercise 4.14, line 2. Insert “relatively prime” after “choose
a, b”.
• page 532, Exercise 25(e), line 10. Change N1 (m) to N1 (2m − 1).
• page 579. The following three items are missing: First edition—Supplementary
Exercise 3.19 = Second edition—Exercise 3.44; First edition—Supplementary
Exercise 3.10 = Second edition—Exercise 3.63; and First edition—
Supplementary Exercise 3.17 = Second edition—Exercise 3.189.