Abstract:It is known that every graph of sufficiently large <a class="topic-link" href="/topics/mathematics/chromatic-number">chromatic number</a> and bounded clique number contains, as an <a class="topic-link" href="/topics/mathematics/induced-subgraph">induced subgraph</a>, a subdivision of any fixed forest, and a subdivision of any fixed cycle. Equivalently, every forest is pervasive, and <math>K3</math> is pervasive, in the class of all graphs, where we say a graph H is "pervasive" (in some class of graphs) if for all <math>ℓ≥1</math>, every graph in the class of bounded clique number and sufficiently large chromatic number has an induced subgraph that is a subdivision of H, in which every edge of H is replaced by a path of at least ℓ edges.Which other graphs are pervasive? It was proved by Chalopin, Esperet, Li and Ossona de Mendez that every such graph is a "forest of lanterns": roughly, every block is a "lantern", a graph obtained from a tree by adding one extra vertex, and there are rules about how the blocks fit together. It is not known whether every forest of lanterns is pervasive in the class of all graphs; but in another paper two of us prove that all "banana trees" are pervasive, that is, <a class="topic-link" href="/topics/mathematics/multigraph">multigraphs</a> obtained from a forest by adding parallel edges, thus generalizing the two results above. This paper contains the first half of the proof, which works for any forest of lanterns, not just for banana trees.Say a class of graphs is "ρ-controlled" if for every graph in the class, its chromatic number is at most some function (determined by the class) of the largest chromatic number of a ρ-ball in the graph. In this paper we prove that for every <math>ρ≥2</math>, and for every ρ-controlled class, every forest of lanterns is pervasive in this class.These results turn out particularly nicely when applied to string graphs. A "chandelier" is a special lantern, a graph obtained from a tree by adding a vertex adjacent to precisely the leaves of the tree. A "string graph" is the <a class="topic-link" href="/topics/mathematics/intersection-graph">intersection graph</a> of a set of curves in the plane. There are string graphs with clique number two and chromatic number arbitrarily large. We prove that the class of string graphs is 2-controlled, and consequently every forest of lanterns is pervasive in this class; but in fact something stronger is true, that every string graph of sufficiently large chromatic number and bounded clique number contains each fixed chandelier as an induced subgraph (not just as a subdivision); and the same for most forests of chandeliers (there is an extra condition on how the blocks are attached together).

Large Induced Forests in Graphs.

Induced forests in some distance-regular graphs

Induced forests in regular graphs with large girth

A Generalization of Boesch's Theorem

A Caro-Wei bound for induced linear forests in graphs

The minimum size of a $3$-connected locally nonforesty graph

Forest Cuts in Sparse Graphs

Induced subgraphs of induced subgraphs of large chromatic number

The maximum size of an induced forest in the binomial random graph

The Turán problem for a family of tight linear forests

Largest common subgraph of two forests

NC Algorithms for Partitioning Sparse Graphs into Induced Forests with an Application

Induced subgraphs density. IV. New graphs with the Erdős-Hajnal property

Induced subgraphs of graphs with large chromatic number. V. Chandeliers and strings

Induced Subgraphs and Tree Decompositions VIII: Excluding a Forest in (Theta, Prism)-Free Graphs

Induced Forests and Trees in Erdős–Rényi Random Graph

Partitioning planar graph of girth 5 into two forests with maximum degree 4

The Strong Nine Dragon Tree Conjecture is true for $d \leq k+1$

Practical Approximation Schemes for Maximum Induced-Subgraph Problems on K_{3, 3}-free or K_5-free Graphs.

Maximal and maximum induced matchings in connected graphs