The annual killifish Austrofundulus limnaeus lives life on the edge. The Venezuelan ponds where it lives rapidly evaporate in the dry season,killing off the adults and exposing their mud-buried embryos to extended periods of drought. Not that drought is the first of the embryo's problems:microbes living in the soil use up all the oxygen. `Not only are they coping with life without water, but also with life without oxygen,' says Jason Podrabsky from Portland State University, who has investigated with his colleagues the root of the embryos' extreme anoxia tolerance(p. 2253).The key to the embryos' survival is that they go into periods of dormancy,or diapause, twice during their development, even if conditions are ideal.`Their environment is very harsh and unpredictable, so diapause occurs as a part of their natural developmental program,' says Podrabsky. Unlike most annual killifish, A. limnaeus embryos skip diapause I at four days of development, but enter diapause II at around 24 days old and diapause III near the end of development. During diapause the metabolism slows down greatly, so could this be the key to surviving anoxia?To assess anoxia tolerance during development, the team collected fertilised embryos from 4 to 32 days old and put them in embryo medium at 25°C before sealing the jars and bubbling nitrogen through them to remove all the oxygen. They then kept a close eye on the developing embryos, opening the jars when they started dying and counting how many of them were still alive. They found that the later embryos were exposed to anoxia, the longer they survived: 4-day-old embryos suffered 50% mortality after 20 days of anoxia, while 32-day-old embryos, in diapause II, suffered 50% mortality after more than 60 days anoxia. This is the longest anoxia survival of any vertebrate at 25°C.The team knew from the brine shrimp, another anoxia tolerant animal that enters diapause, that there is a window of opportunity after diapause where the animal can shut down again even if it has restarted its development. To find out if this window existed in the killifish embryos, they collected embryos just out of diapause II and tested their anoxia tolerance. They found that embryos exposed to anoxia 4 days after the end of diapause were still very tolerant, but lost this tolerance when they were tested 8 days after,showing that post-diapause safety cushion in the killifish is about 4 days.But what about changes in metabolism? The team took embryos of different developmental stages and flash froze them to preserve their metabolites. After mashing them up and extracting their metabolites, which they analysed using chromatrography and mass spectrometry, they found that anoxic embryos produced lactate, succinate and the amino acid alanine, which are end-products of anaerobic metabolism. This suggests that the embryos survive by relying on glycogen stores and reducing their metabolic rate, but that they haven't developed an alternative strategy for dealing with anoxia. Embryos that accumulated lactate at the slowest rate also survived the longest.They also found that anoxic embryos produced lots of the neurotransmitter GABA, which is known to protect neurons in other anoxia tolerant species, such as turtles and the crucian carp. Podrabsky explains that it's possible that GABA is protecting the traditionally anoxia-sensitive brain and heart tissue in diapause II embryos. For now, the killifish embryo can elbow its way onto the roll-call of champion vertebrate anoxia survivors.