Two stars are better than one Fred Cline, MWSC student Jeffrey L. Poet, faculty project mentor Missouri MAA, St. Joseph April 1, 2005 In 2002, Sin-Min Lee and Man Kong 1 conjectured that every constellation graph with an odd number of stars is super edge-magic. Constellations with three stars were investigated by a group of MWSC student researchers. 2 This paper represents progress toward classifying which constellation graphs with two stars are super edge-magic. We begin with some definitions. First, an n-star is a graph on n 1 vertices such that one vertex, called the central vertex, is connected to each of the other vertices with a single edge and these are the only edges in the graph. Examples of a 3-star and a 4-star are given below. Second, a constellation is a union of star graphs. We denote the constellation with two stars, an n-star and an m-star where n m , by the ordered pair (n, m) . As an example the constellation (3, 4) is given below. Third, a graph with v vertices and e edges is called super edge-magic if the vertices can be labeled with the integers 1, 2, integers v 1, v 2, ,v and the edges can be labeled with the , v e such that the sum of the labels on any edge and its two endpoints is the same regardless of which edge is chosen. The example below shows a super edge-magic labeling of the constellation 3,4 . Note that for any edge, the sum of the label on that edge and the labels on its two endpoints is 21. Lee and Kong 1 showed that for any positive integers n and k, the constellation n, k (n 1) is super edge-magic. We provide evidence here to support our conjecture that these are the only constellations with two stars that are super edge-magic. We do not yet have a complete proof, but hope to soon. To simplify our presentation, we next introduce the idea of a vertex-sum. For each edge in a labeled graph, we define the vertex-sum of the edge to be the sum of the labels of the two endpoints of the edge. Observe that a graph is super edge-magic if and only if the set of vertex-sums of the graph is a set of consecutive integers. For the labeled constellation (3,4) above, the set of vertex-sums is 6,8,10,5,7,9,11 . The purpose of this paper is to show that constellations of the form 4N 3,4M 1 are not super edge-magic for any positive integer values of N and M. To do so, we assume an appropriate labeling of the vertices exists and then derive a contradiction. To this end, suppose that a super edge-magic labeling of 4 N 3,4M 1 exists and that the central vertex of the 4 N 3 -star is labeled k and the central vertex of the 4M 1 -star is labeled . Then the total of the vertex-sums of the graph is Total of vertex-sums 1 2 4 N 3 4 M 1 2 4 N 4 k 4 M 2 . where the first part of the sum is the sum of each of the vertex labels one time, the second part accounts for the use of k as an endpoint of 4N 3 edges (but one use has been included in the first part of the sum), and the last part accounts for the multiple use of the label in a similar way. Note that the first part of the above computation is the sum of the first 4N 4M 2 positive integers. We now use a familiar formula to rewrite this sum and we consider the parity of our sum. Total of vertex-sums 4 N 4M 2 4 N 4M 1 2 4 N 4 k 4M 2 2 N 2M 1 4 N 4M 1 4 N 4 k 4M 2 . odd even even odd We next work to derive a contradiction. Since the graph is assumed to be super edge-magic and the given labeling is assumed to illustrate this fact, we know that the set of 4 N 3 4M 1 4 N M 1 vertex-sums form a set of consecutive integers. Observe that the sum of any four consecutive integers is even, as illustrated by p p 1 p 2 p 3 4 p 6 . It follows that the sum of any set of consecutive integers with some multiple of 4 elements must be even. This contradicts that the total of the vertex-sums is odd. Therefore, any constellation of the form 4 N 3,4M 1 is not super edge-magic. Note that the argument also proves that constellations of the form 4M 1,4 N 3 are not super edge-magic because the fact that 4N 3 4M 1 was never used in the argument. Why are constellations with two stars better than those with one star? It is trivial that a single n-star is super edge-magic as shown below. On the other hand, some constellations with two stars are super edge-magic and others are not, making constellations with two stars far more interesting. (1) S.-M. Lee & M. Kong, On super-edge magic n-stars, Journal of Combinatorial Mathematics and Combinatorial Computing, 40(2002), 87-96. (2) J. Poet, V. Onkoba, D. Daffron, H. Goforth, & C. Thomas, On super-edge magic labelings of unions of star graphs, Journal of Combinatorial Mathematics and Combinatorial Computing, 53(2005), 49-63.