Detailed information available in the Supplementary Methods and Materials. Caenorhabditis elegans strain Bristol N2 was grown on NGM at 20°C with E. coli OP50 as food source30 (link). Morphological changes in worms between the age of 48 and 67 hours after synchronisation were detected by analysing pictures of individual worms that were taken at a magnification of 100×. Samples for microarray analysis were drawn every hour between 44 and 58 hours after synchronisation by bleaching. RNA was isolated with either a RNEasy Micro Kit from Qiagen or with use of a Maxwell® 16 AS2000 instrument with a Maxwell® 16 LEV simplyRNA Tissue Kit (both Promega). After RNA isolation, one, two or three independent replicates per time point were analysed with microarrays (C. elegans (V2) Gene Expression Micorarray 4X44K slides from Agilent) following the ‘Two-Color Microarray-Based Gene Expression Analysis’-protocol from Agilent. Data were extracted with the Agilent Feature Extraction Software. All statistical analyses were performed in the ‘R’ statistical programming software (version 2.13.1 × 64). For normalization the Limma package was used with for Agilent array recommended settings31 (link). For k-means clustering, 20 iterations on 10 different starting situations were performed. Twelve k-means clusters were formed to be able to visualise gene expression changes properly. Enrichment studies were done using a hyper-geometric test. The physical distance from a gene to its closest neighbour within its k-means cluster was measured as the distance from start-site to start-site. The mutation frequency19 (link) of the genes from the k-means clusters was calculated using a permutation analysis. Heat-maps were constructed with the ‘heatmap’ function. Data of published gene expression studies were obtained from SPELL24 (link). Bacteria and genotypes for the bacterial food experiment are from20 (link). Data was stored in WormQTL32 (link)33 (www.WormQTL.org).