---

Claim. For every $s\ge2$ the rational $\alpha=\frac32-\frac1{2s}$ is a Turán exponent, and for all integers $s,k\ge1$ so is $1+\frac{s}{sk+1}$. Writing $K'_{s,t}$ for the $1$-subdivision of $K_{s,t}$, the paper proves $\mathrm{ex}(n,K'_{s,t})=O(n^{3/2-1/(2s)})$ for integers $2\le s\le t$ (Theorem 1.8 of arXiv:1903.10631v2), a conjecture of Kang, Kim and Liu, and this is tight up to the constant once $t$ is large in terms of $s$; it also exhibits, for all $s,k\ge1$, a bipartite graph $L$ with $\mathrm{ex}(n,L)=\Theta(n^{1+s/(sk+1)})$, which makes $1+1/k$ a limit point of the Turán exponents for every $k$. The statements are recorded on the library's source card.

Covers. The instances $\alpha=\frac32-\frac1{2s}$ for $s\ge2$ (the family the site's commentary lists for [CJL21]) and $\alpha=1+\frac{s}{sk+1}$ for $s,k\ge1$, each realized by a single bipartite graph. The statement for every rational $\alpha\in[1,2)$ is settled by the accepted claim page Adamczewski 2026.

Depends on. Nothing in this wiki; the result is the paper's own theorem.

Acceptance. Refereed: D. Conlon, O. Janzer and J. Lee, More on the extremal number of subdivisions, Combinatorica 41 (2021), no. 4, 465--494, doi:10.1007/s00493-020-4202-1, a refereed journal. First posting: arXiv:1903.10631, v1 25 March 2019 (the date this page is named by), v2 25 April 2020. No reviewed evidence is listed: the site's commentary lists these exponents among the Turán exponents known before 2026 and credits the paper, but its label credits GPT-6 Astra with the full proof and is not an acceptance of this result.

Read depth. The statements are taken from the paper's abstract and the result list on the library card; no proof was read, and nothing is independently reviewed in this corpus.