Chromosomal aberrations are often listed as a significant cause of early embryonic death in the mare, despite the absence of any concrete evidence for their involvement. The current study aimed to validate fluorescent in situ hybridization (FISH) probes to label specific equine chromosomes (ECA2 and ECA4) in interphase nuclei and thereby determine whether numerical chromosome abnormalities occur in horse embryos produced either in vivo (n = 22) or in vitro (IVP: n = 20). Overall, 75% of 36,720 and 88% of 2,978 nuclei in the in vivo developed and IVP embryos were analyzable. Using a scoring system in which extra FISH signals were taken to indicate increases in ploidy and "missing" signals were assumed to be "false negatives," 98% of the cells were scored as diploid and the majority of embryos (30/42: 71%) were classified as exclusively diploid. However, one IVP embryo was recorded as entirely triploid and a further seven IVP and four in vivo embryos were classified as mosaics containing diploid and polyploid cells, such that the incidence of apparently mixoploid embryos tended to be higher for IVP than in vivo embryos (P = 0.118). When the number of FISH signals per nucleus was examined in more detail for 11 of the embryos, the classification as diploid or polyploid was largely supported because 2,174 of 2,274 nuclei (95.6%) contained equal numbers of signals for the two chromosomes. However, the remaining 100 cells (4.4%) had an uneven number of chromosomes and, while it is probable that many were artefacts of the FISH procedure, it is also likely that a proportion were the result of other types of aneuploidy (e.g., trisomy, monosomy, or nullisomy). These results demonstrate that chromosomally abnormal cells are present in morphologically normal equine conceptuses and suggest that IVP may increase their likelihood. Definitive distinction between polyploidy, aneuploidy and FISH artefacts would require the use of more than one probe per chromosome and/or probes for more than two chromosomes.