Embryos from vastly different vertebrates, such as fish, chickens, and humans, look strikingly similar during a specific window of development because they share a common genetic toolkit for building a body. This similarity peaks during a mid-embryonic phase called the pharyngula stage, when all vertebrate embryos display the same basic architecture: gill-like arches, a rod-shaped support structure called a notochord, a spinal cord, and primitive kidneys. Before and after this stage, embryos actually look quite different from one another. Understanding why requires looking at both the shared genes and the evolutionary pressures that keep this middle phase locked in place.
The Hourglass Shape of Development
If you compare vertebrate embryos across their entire development, the pattern of similarity isn’t a straight line. It’s shaped like an hourglass. Early development varies widely between species, the middle narrows into a period of striking resemblance, and then late development diverges again as species take on their adult forms. This is known as the developmental hourglass model.
The narrow waist of the hourglass is the phylotypic period, roughly corresponding to the pharyngula stage. A 2011 study in Nature Communications confirmed this by comparing the active genes in mouse, chicken, frog, and zebrafish embryos. Gene activity was most similar across all four species during the pharyngula stage, while earlier and later stages showed much more divergence. In other words, these animals don’t just look alike at this stage. Their cells are reading remarkably similar genetic instructions.
The reason early development differs so much is surprisingly practical. A frog egg is packed with yolk in a way that forces cells to divide completely around it (holoblastic cleavage), while a chicken egg has so much yolk that cell division only happens in a small disc on top (meroblastic cleavage). Mammalian eggs have almost no yolk at all and divide in a unique rotational pattern. These physical differences in egg structure drive very different early cell division strategies. Yet all these different starting points funnel into the same basic body plan by mid-development.
A Shared Genetic Toolkit
The most important set of shared instructions comes from a family of genes called Hox genes. These genes act as master switches that tell cells where they are along the head-to-tail axis of the body. What makes them remarkable is that they’re arranged on the chromosome in the same order as the body regions they control: genes at one end of the cluster shape the head, genes in the middle shape the torso, and genes at the other end shape the tail. This mirror-like arrangement, called colinearity, is conserved across virtually all animals with bilateral symmetry, from insects to humans.
Hox genes appeared early in animal evolution, after sponges branched off but before the split between jellyfish and everything else. In jellyfish and corals, these genes are scattered randomly across the genome, which may be one reason those animals have radial symmetry rather than a distinct head and tail. In all bilaterally symmetric animals, Hox genes are clustered together and activated in precise spatial and temporal sequences. This shared arrangement is a major reason why a developing fish and a developing human go through a phase where their body plans look nearly identical: the same set of genetic instructions is laying out the same fundamental architecture.
The vertebrate skeleton provides a clear example. Hox genes subdivide the spine into cervical, thoracic, lumbar, sacral, and caudal regions from front to back. The specific number of vertebrae in each region varies between species, but the overall organizational logic is the same because the same genes are doing the work.
Why Evolution Can’t Easily Change This Stage
The phylotypic period stays conserved because the genes active during this stage are doing too many things at once. A single gene active during mid-embryonic development might simultaneously influence the nervous system, the skeleton, and the kidneys. This property, called pleiotropy, means that a mutation affecting one of these roles would likely disrupt the others, producing an embryo that can’t survive. Researchers at Genome Research have described the phylotypic period as an “evolutionary regulatory lockdown.”
This lockdown operates through multiple mechanisms. Strong evolutionary pressure weeds out harmful mutations in genes active at this stage. But there’s also evidence that even beneficial mutations are less common during this period. The genetic switches (called enhancers) that control gene activity during the phylotypic stage show signs of reduced positive selection, meaning evolution isn’t just blocking bad changes but also generating fewer new ones. The combination of these forces, along with a degree of built-in genetic robustness that buffers against small mutations, keeps this developmental window remarkably stable across hundreds of millions of years of evolution.
Shared Structures With Different Fates
The most visible sign of embryonic similarity is the set of pharyngeal arches, the bulging structures along the neck region that all vertebrate embryos share. In fish, the third through seventh arches become gills. In humans, they become the jaw, the bones of the middle ear, and the muscles and blood vessels of the face and neck. A piece of cartilage called Meckel’s cartilage forms part of the lower jaw in fish but becomes the malleus, one of the tiny bones of the middle ear, in mammals. Similarly, a bone called the hyomandibular that suspends the jaw in fish has been repurposed as the stapes, another middle ear bone, in mammals.
The pharyngeal arches and their segmental arrangement are highly conserved from invertebrate chordates like the lancelet all the way through birds, reptiles, and mammals. This is why a human embryo at five weeks has structures that look like the beginnings of gills. They aren’t gills and never will be, but they’re built from the same developmental program that produces gills in fish. Evolution didn’t invent new structures from scratch for each lineage. It repurposed the same embryonic building blocks.
The Human Embryonic Tail
Human embryos also develop a visible tail, another feature that underscores their resemblance to other vertebrate embryos. The tail grows as paired segments of tissue called somites form along the body axis. Caudal (tail) somites increase steadily, peaking at 13 to 16 pairs around Carnegie stage 16, which corresponds to roughly 37 to 42 days after fertilization. Between Carnegie stages 16 and 17, about five pairs of caudal somites disappear in a dramatic reduction driven by programmed cell death. Immune cells called macrophages actively clear the dying tissue. After this rapid regression, only 7 to 11 pairs of caudal somites remain, eventually forming the coccyx, the small fused bones at the base of your spine.
Haeckel’s Drawings and What They Got Wrong
The idea that embryos from different species look alike became famous in the 1870s through illustrations by the German biologist Ernst Haeckel. His drawings showed embryos of fish, salamanders, turtles, chickens, rabbits, and humans lined up in neat rows, looking almost identical in the earliest stages and gradually diverging. The concept was catchy, often summarized as “ontogeny recapitulates phylogeny,” meaning that an organism’s development replays its evolutionary history.
The problem is that Haeckel took considerable artistic license. Comparisons by the embryologist Michael Richardson and colleagues showed that Haeckel’s “earliest” embryos were actually from a midpoint in development, not the true beginning. He also removed the yolk from his illustrations, which made embryos from different species appear more uniform than they really are. The yolk distorts the shape and posture of embryos in ways that have nothing to do with evolution or development. Early-stage embryos are, in reality, quite different from one another due to the variation in egg size, yolk content, and cleavage patterns described above.
The core observation, that vertebrate embryos pass through a stage of genuine similarity, holds up under modern analysis. But the similarity is concentrated in mid-development, not at the very beginning, and it’s less uniform than Haeckel’s drawings suggested. The hourglass model, backed by gene expression data from multiple species, provides a far more accurate picture than the old idea that development simply replays evolution in order.