Is Palaeolithic cave art consistent with costly signalling theory? Lascaux as a test case