A red salamander with black spots holds down a pile of papers on his desk, and yellow and green frogs peer over his shoulder from a bookcase. They sit rather casually for frogs, with legs dangling like humans' over the shelf edge. Also seated in this office is paleontologist Neil Shubin. He wears an easy, almost impish smile, and if you get him talking about frogs or fossils or fish fingers--admittedly, not a hard thing to do--you'll soon find that his eyes smile too. Shubin is interested in how things like fins, wings, and fingers came into being. With a Guggenheim Fellowship, he is on sabbatical this year.

At home in the laboratory and in the field, Shubin brings together two scholarship methods to study the evolution of body structures: genetic and embryonic research, and fossil hunting and analysis. The study of gene expression--what genes are "turned on" and when--in the embryos of mammals, birds, fish, insects, and crustaceans yields a rich store of comparative data that contributes to understanding the genetic engine driving evolution. "We look at the fossils of extinct forms," he explains, "but we also do comparative studies of genetic and developmental systems of existing forms. What's interesting is that we can use developmental genetics to make predictions about what we should see in the fossil record. Then we go to the fossil record and test those predictions."

The great diversity of living creatures adapted to air, water, and land by the invention of new body features, particularly new kinds of specialized appendages, that helped them survive in particular environments. Collaborative investigations with geneticists have led Shubin to theorize that an ancient and deep genetic relationship underlies the multiformity of animal limbs. Genetic researchers and developmental biologists have found that similar genes get turned on in the development of embryo limb buds across different species. These similarities have been retained, despite half a billion years of independent evolution. "The same genes are doing different things in today's species," argues Shubin making the leap from living animals to fossils. "So it's clear that one of the things that happened in the long process of evolution, which led to the origins of fingers and toes, was that nature used old genes in new ways."

A key step in the evolution of animal limbs, says Shubin, was the establishment of a mechanism that initiated and patterned the formation of outgrowths from the body wall. This ancient genetic system provided a foundational scheme whose modification and redeployment both promoted and constrained the evolution of structures as diverse as fins, wings, legs, antennae, and arms. "Although their respective evolutionary histories are unique," he writes with his collaborators, "vertebrate, insect, and other animal appendages are organized by a similar genetic regulatory system that may have been established in a common ancestor" perhaps as long ago as 600 million years. Natural selection, imposed externally upon species by interaction with their environment, is not the only engine driving evolution, he argues. Genetic structures also govern change from within.

To help understand how new body features evolved, part of Shubin's research program is to uncover fossils left behind by the earliest representatives of existing or extinct species. He has hunted fossils in the deserts of the American Southwest and on the frozen tundras of Greenland. These regions hold rock formations from the Triassic Period, a time when the first dinosaurs, mammals, frogs, and salamanders appeared over 200 million years ago. Another reason he hunts in such forbidding regions is because little excavation is needed to find fossils. The bedrock, a repository for many fossilized specimens, has been exposed by the erosion characteristic of these harsh climates, leaving the bones and dinosaur footprints right on the surface.

As a graduate student at Harvard, he was part of an expedition that discovered the earliest known frog, Prosalirus bitis, in northern Arizona. The find actually consisted of remnants from four frogs. "It was in really bad shape," he says. "It looked like roadkill." Shubin believes Prosalirus descended from a salamander-like amphibian with a long tail. As that creature evolved, it developed long and powerful hind legs and an elongated pelvis that connected to the spine by a movable joint. As the tail grew shorter, its bone segments receded into the pelvis, eventually fusing into a solid rod called the urostyle. That structure anchors the muscles that formerly moved the tail. When the frog jumps, the hinged pelvis allows the creature to lift and align its body along the trajectory of the leap. The stability formerly provided by a fused pelvis and spine was taken over by the muscles secured to the urostyle. "So what you've done," explains Shubin, "is make a frog. You've taken the muscles that formerly wagged the tail and changed their function. So now, instead of moving the tail during a walk, the muscles provide stability during a jump by fastening the remnants of the tail skeleton to the pelvis." Inferences from how existing frog mechanisms function help Shubin understand what happened when Prosalirus hopped among the dinosaurs. He consults the fossil record to piece together how the ancient frog's Triassic mechanisms and body plan came into being.

In another place where Shubin scavenges for fossils, they aren't lying around like a paleontologist's smorgasbord. Still, he doesn't need to dig. The Pennsylvania Department of Transportation, using some of its biggest equipment to cut through mountains and open up roads, allows Shubin on its excavation sites upstate to pick over the entrails of mountains. "They're exposing a unique window on evolutionary history with virtually every scoop of rock they take," he notes. The Catskill Formation, which runs through much of northern Pennsylvania not far from Penn, contains layers of red rock deposited nearly 400 million years ago in the Devonian Period. At that time, the state consisted mainly of muddy streams and swamps teaming with insects and primitive plants. The first amphibians were still millions of years from crawling out of the water, but there were plenty of fish. Among them were large bottom-dwelling carnivores with heavy jaws and huge teeth. Shubin sketches an arms-race scenario where the fish were getting bigger or staying small but evolving formidable body armor. "It was a hostile environment," he says, "a real fish-eat-fish world."

Scavenging among the talus of a road cut where PennDOT was putting a new bend in Route 15 in north-central Pennsylvania, Shubin and colleague Ted Daeschler, Gr'98, made a discovery that has overturned one of the basic assumptions of evolutionary theory. Daeschler is a curator at Philadelphia's Academy of Natural Sciences. About 20 miles from that site, they had previously discovered the fossil of an amphibian estimated to be 365 million years old, the oldest tetrapod ever found in North America. More than just being old, the discovery indicates that vertebrates adapted to land earlier than previously thought. But their new discovery would challenge more than a few points of chronology in the history of evolution.

The new fossil was about 370 million years old and turned out to be a lobe-shaped fin from the right front side of a six to eight foot fish. The unusual thing was that the bone structure, although unmistakably from the shoulder of a fish, did not resemble the skeleton of a fin. The fan-shaped array of bones revealed that the fish possessed a limb of unusual dexterity--one that could adjust and conform its shape in subtle ways, more like fingers than a fin. "We were dumbfounded," declares Shubin. "Essentially, this is the same pattern we have in our limbs. And if you look at the number of rods or digits this fish had--eight--it's the same number of digits that the earliest amphibians had."

Shubin is quick to point out that the fin evolved for use in water. The paper-thin bones were not meant for walking on land; they were not strong enough to bear the load of the fish's body. What function the fish fingers did serve is not clear. Shubin speculates the fin may have been used to maneuver in shallow water, perhaps to pull the creature along the bottom or through weed-choked sections of the streams. Although not a direct ancestor of the first amphibians, he believes his fingered fish is a close cousin. In the Devonian arms race, instead of getting bigger or growing armor, one strategy for coping could have been to simply leave the battle--crawl out of the streams inhabited by the giant carnivore fish. "Clearly, one factor here has to be predatory escape," notes Shubin. "So, it could be that one lineage hit on this bony system to maneuver in water. But also, after some small modifications, the creature could function on land, and then the new structure was exploited to escape."

Before the unearthing of fish fingers, scientists believed that species evolved new structures as a way of adapting to new environments. The story of evolution had said that, after the first amphibian-fish pulled itself, awkward and gasping, out of the water, its successors developed legs for walking around on land. With Shubin and Daeschler's discovery, the fossil record tells a different story. Limbs for locomotion didn't first appear on land; they began to evolve in water. When they crawled onto the land, those first terrestrial creatures didn't adapt to a new environment: they were already prepared for it. "Here's what's interesting about the fossil," says Shubin. "When you're inventing new things in evolution, part of what's going on is that these animals are using old genes in new ways. Evolution doesn't wait for new genes at every step. Fingers and toes did not evolve to walk on land; they evolved initially to do something in water. So, part of the evolution of fingers and toes involved a change in function, finding new functions for the structures that already existed. A lot of the evolution of new structures involved subtle changes in function from what already existed--using old things in new ways."

Using old things in new ways--Neil Shubin's scholarship is like that too. Fossils and genetics: something old, something new. A paleontologist, he crosses into the discipline of embryology: that's something borrowed. But his eyes are smiling now: there's nothing blue.