<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<item>
  <id>05565365</id>
  <dt>j</dt>
  <an>05565365</an>
  <augroup>
    <au>Gy\'arf\'as, Andr\'as</au>
    <au>S\'ark\"ozy, G\'abor N.</au>
    <au>Schelp, Richard H.</au>
  </augroup>
  <ti>Multipartite Ramsey numbers for odd cycles.</ti>
  <so>J. Graph Theory 61, No. 1, 12-21 (2009).</so>
  <py>2009</py>
  <pu>John Wiley \& Sons, New York, NY</pu>
  <lagroup>
    <la>EN</la>
  </lagroup>
  <ccgroup>
  </ccgroup>
  <utgroup>
    <ut>Ramsey numbers</ut>
    <ut>regularity lemma</ut>
    <ut>complete 5-partite graph</ut>
    <ut>monochromatic cycle</ut>
  </utgroup>
  <cigroup>
  </cigroup>
  <ligroup>
    <li>doi:10.1002/jgt.20364</li>
  </ligroup>
  <abgroup>
    <ab>The Ramsey number $r(C_n,C_n) = 2n - 1$ for $n$ odd. The authors conjecture the stronger result that if the edges of a complete $5$-partite graph $$ K_{\frac{n-1}{2},\frac{n-1}{2},\frac{n-1}{2},\frac{n-1}{2},1} $$ are $2$-colored, then there will be a monochromatic $C_n$. Thus, deleting the edges from $4$ vertex disjoint copies of a $K_{(n-1)/2}$ from the complete graph $K_{2n - 1}$ will still result in a monochromatic $C_n$. They show that if the edges from a $K_{(n+1)/1}$ are deleted from a $K_{2n-1}$, then there is a $2$-coloring without a monochromatic $C_n$. To support the conjecture they show that for any $\epsilon > 0$ and $n$ sufficiently large, any $2$-edge coloring of a complete $5$-partite graph $K_{m_1,m_2,m_3,m_4,m_5}$ with $\epsilon n \leq m_1 \leq m_2 \leq m_3 \leq m_4 \leq m_5 \leq n/2$ and $m_1 + m_2 + m_3 + m_4 + m_5 = (2 + \epsilon)n$ will contain a monochromatic $C_n$.</ab>
    <rv>Ralph Faudree (Memphis)</rv>
  </abgroup>
</item>