One day, I took a walk up a canyon in the El Paso Mountains, which run northeast from Red Rock Canyon in the Mojave. The El Pasos have produced as wide a variety of fossils as any part of the California desert—not only the Pliocene mammals that John Merriam found in the Ricardo Formation, but petrified wood later dug from the same formation as well as older animal and plant remains from others. I didn’t find any fossils that day, but I did encounter one of California’s most influential fossil collectors.
The porous rock of the canyon was eroded into many little caves. Peering into one of these piled with sticks and leaves, I was startled to see a rabbit skull staring back at me, evidently placed there by a desert wood rat (Neotoma lepida) that had gathered the sticks to make its den. Wood rats are also called pack rats because of their collecting habits. I made squeaking sounds to try to get the occupant’s attention, without much anticipation. I’ve squeaked outside a lot of dusky-footed wood rat (N. fuscipes) houses and been ignored. That species, which inhabits dense coastal vegetation, seldom comes out in daylight.
But desert wood rats are more confiding, perhaps because of their lonelier environment. A plump, big-eared individual popped out of the cave and eyed me, disappeared, and then emerged again for a closer look before it had satisfied its curiosity. Although not as furry as a mountain relative, the bushytailed wood rat (N. cinerea), it was a pretty mammal—more like a pika than a Norway rat. I continued up the canyon, and when I later stopped to rest on a pile of boulders, I squeaked again. There was no sign of a den nearby, but another wood rat, smaller than the first, peered at me from another cave. I suppose I could have spent the day wasting curious wood rats’ time.
The California desert is literally stuffed with wood rat dens because there are a lot of wood rats and because their dens last a long time. They contain everything that their occupants have collected in their lives, also everything they have excreted. Wood rats get all the water they need by eating green plant material, but they don’t metabolize it from dry seeds and concentrate their urine as kangaroo rats do, so, like water drinkers, they urinate a lot. The urine dries into an amberlike substance that binds twigs, leaves, and feces—sometimes bones, feathers, and other oddities—into a deposit called a midden. Because of the urine’s chemistry, middens are resistant to decay.
Created by generations of wood rats, old middens can fill caves and are curious objects. I’ve seen small ones that resembled slabs of peanut brittle, and large ones can be surreal looking. It took white explorers awhile to figure them out. William Manly described a possible encounter with one near Death Valley in 1849:
So we turned up a cañon leading to a mountain and had a pretty heavy upgrade and a rough bed for a road. Partway up we came to a high cliff and in its face were niches or cavities as large as a barrel or larger, and in some of them we found balls of a glistening substance looking somewhat like pieces of variegated candy stuck together. The balls were as large as small pumpkins. It was evidently food of some sort, and we found it sweet but sickish, and those who were so hungry as to break off one of the balls and divide it among the others, making a good meal of it, were a little troubled with nausea afterwards.
The party suspected they were plundering an Indian cache as they’d done earlier with some squashes, an expedient that bothered the scrupulous Manly. One historian thought Indians might have made the glistening balls by processing reed stems into a kind of syrup. Later observers were convinced that they were from wood rat middens, in which case Manly needn’t have worried. The Indians knew about those and probably would have been amused to see the greedy invaders eating dried rat urine.
In the 1870s, Edward Cope seems to have been the first scientist to guess what the curious objects are. Some biologists had thought they were deposited by chuckwallas, which, along with a variety of other creatures, do sometimes live in large wood rat dens. Another den lodger, the desert iguana, is known to supplement its diet with wood rat feces, perhaps as a way of acquiring digestive microbes along with extra nourishment. Herpetologists speculate that this may have helped the lizards to become herbivores, since the microbes break down tough cellulose. But herbivorous lizard excrement somehow doesn’t have what it takes to make massive, durable deposits.
Wood rat middens eventually afforded a more edifying use than the forty-niners’ impromptu snack. Another desert mystery that emerged from early evolutionary studies was that fossils revealed less about vegetation in the Pleistocene ice age than in the older Tertiary period. Deposits of fossil leaves, fruits, and branches such as Daniel Axelrod dug from Miocene and Pliocene strata are scarcer in later Pleistocene ones. The Pleistocene was shorter than the previous epochs so fossil beds had less time to form, and rapid drying of lakes during arid interglacials would have worked against fossilization of plant deposits. Whatever the reasons, the main known sources of Pleistocene plant remains were pollen deposits in various places and ground-sloth dung found in a few caves. Since wind carries pollen long distances and the sloths wandered widely, neither source was reliable for determining a given location’s former vegetation.
Passage through a giant mammal’s guts can cause plant identification problems as well. When sloth dung first turned up in the Mojave in the 1930s, Forrest Shreve thought the material in it was from nondesert plants. But Philip Munz, a younger botanist rival of the great plant describer Willis Jepson, identified it as originating from Joshua tree and the saltbush species called desert holly. Other dung deposits, at most around twelve thousand years old, suggested that more trees and fewer desert shrubs had occupied the region in the Pleistocene, but there wasn’t enough to be sure.
The situation changed in the 1950s when the U.S. government spread a lot of research money around the desert in promoting its thermonuclear agenda. Two of the scientists who pursued this largesse into the Mojave were Phillip V. Wells, a botanist, and Clive D. Jorgenson, a zoologist. Interested in the Pleistocene vegetation mystery, they intended to visit a sloth cave on a 1961 expedition, but first made a side trip to see if juniper trees, living relicts of ice age climate, were still growing on top of a certain mountain. Finding no junipers, they stopped to rest in a canyon: “As they sat and commiserated, Jorgensen looked over his shoulder and something caught his eye about thirty meters away—a dark, shiny mass beneath an overhang. He walked over to it, grabbed a chunk, and broke it open, calling out to Wells: ‘You’ve got to see this! This is where all the junipers are!’”
Three years later, Wells and Jorgensen published an article in Science predicting that radiocarbon dating of rat midden plant remains would help to solve the Pleistocene vegetation mystery: “The limited foraging range of the wood rat assures that macrofossils preserved in their ancient middens represent relatively local vegetation. Therefore, the middens probably contain more precise information on local paleoclimate than sediments yielding fossil pollen because of the wide dissemination of many types of pollen, especially that of wind-pollinated conifers. Hence, Neotoma middens may have unique value.”
In 1966, Wells and Rainer Berger, an anthropologist, published another Science article fulfilling the prediction: “Seventeen ancient wood rat middens, ranging in radiocarbon age from 7,400 to 19,500 years and to older than 40,000 years, have been uncovered in the northeastern, north-central, southeastern, and southwestern sections of the Mojave desert. Excellent preservation of macroscopic plant materials (including stems, buds, leaves, fruits and seeds) enables identification of many plant species growing within the limited foraging range of the sedentary wood rat.” The material showed that, before around nine thousand years ago, southeast California’s vegetation was very different from now. Even though the fossil middens were in present desert areas, most contained remnants of juniper, and some contained pinyon pine, mountain mahogany, snowberry, ceanothus, and other plants that now occur only in higher elevation nondesert in the Mojave. A few even contained maple and ash tree remains. “Prevalence of woodland vegetation at moderately low elevations throughout the Mojave Desert as recently as about 9,000 years ago is apparent,” Wells and Berger concluded.
The only place in the Mojave where middens from before nine thousand years ago contained desert plants was Death Valley, and even those—from almost twenty thousand years ago—were plants that grow in moister climates than Death Valley’s now. In the Sonoran Desert, far south into Baja, fossil middens also contained quite different plants from today’s, including juniper, although some of them were desert plants like sagebrush and Joshua trees that now grow farther north.
The furry paleontologists had provided a striking corroboration of Axelrod’s prediction two decades earlier that today’s deserts would prove to be a very recent biome, and he was quick to acknowledge this. “That regional climate was not desert, as it is today, is apparent,” he exulted. “The regional occurrence of juniper woodland is demonstrated by megafossil remains preserved in ancient woodrat middens at approximately twenty locations in the Mojave and adjacent desert areas.”
Wood rat paleontologists seemed made-to-order for Axelrod’s version of desert evolution, and not just because of their midden collections. According to the scanty fossil record, the wood rat genus, Neotoma, originated from a group of rodents that arrived in North America from Eurasia in the early Miocene epoch and then underwent “a modest radiation in the late Miocene, some 7 million years ago.” The wood rats now inhabiting the arid West may have evolved even later, perhaps less than a million years ago.
Wood rats seem exemplary of G. Ledyard Stebbins’s theory of accelerated evolution in dry areas. Today two wood rat species, N. floridana and N. magister, inhabit the United States east of the Mississippi, while some twenty species live in the West, Mexico, and Central America. The California desert region has two arid-adapted species—the desert and white-throated (N. albigula)—while two others—the dusky-footed and bushytailed—live on its margins. This suggested that western wood rats began to speciate more rapidly as mountains rose and dryness increased. In accordance with Stebbins’s three conditions for quantum evolution, habitats would have become more diverse, wood rats would have been more dispersed, and their ability to build dens and get their water by eating plants (even dusky-footed wood rats in riparian swamps can do so) would have preadapted them to desert life.
Wood rats may be so preadapted to desert life that desert species do better than woodland ones, which suffer high mortality during the wet winters of the eastern United States and Pacific coast. The eastern wood rats are rare or endangered in some of their range. Desert wood rats seem almost invulnerable in their massive dens, which they often fortify with cactus—also one of their main food items. They seem the opposite of Darwin’s idea that desert organisms mainly struggle with their environment. Desert-dwelling wood rats may struggle mainly with other wood rats, since they are solitary and new generations are always trying to occupy good den sites.
Of course, wood rat middens did not entirely elucidate even the two million years of Pleistocene evolution. Because of the limitations of radiocarbon dating, fossil middens didn’t provide information about vegetation before about fifty thousand years ago. They still didn’t show how or when desert plants had appeared in California. In some cases, they complicated matters.
Creosote bush is an example. As Axelrod’s original 1950 paper on desert evolution acknowledged, its presence in both South and North America implies “long distance migration.” South America has four creosote bush species to North America’s one, and two of the South American species live in thorn scrub and dry woodland as well as in desert. Those species are leafier than desert species. This suggests that Larrea first evolved from nondesert vegetation in South America and that North America’s one desert-adapted species got here much later. But how did it arrive?
When scientists began dating wood rat middens, creosote bush remains appeared only in those of the past seven thousand years. This abrupt appearance did not provide much time for the species to have spread north by the gradual dispersal by which plants usually “migrate.” It suggested some more drastic mode of transportation. One migratory possibility, originally raised by Darwin, was that birds might carry seeds on their way from one continent to another during their seasonal passages. Darwin did experiments to show that seeds can cling to birds’ feet and remain viable in their guts. Axelrod and Stebbins thought this a possible cause for the apparently sudden arrival of creosote bush and other desert plants with puzzling distributions. But there was no evidence for it.
The mystery cleared up slightly when creosote bush remains over ten thousand years old appeared in an Arizona wood rat midden. The species’ arrival in North America had been less abrupt than it seemed, and the genetics of its distribution here support this. Creosote bush grows in the Chihuahuan Desert as well as the Sonoran and Mojave, and its chromosome count changes from east to west. Chihuahuan bushes are mostly diploid; Sonoran ones are mostly tetraploid; Mojave ones mostly hexaploid. This suggests that the species has migrated from east to west with chromosomes multiplying as it colonized new regions, perhaps in response to conditions new to the original population. As Axelrod had noted with California desert wildflowers, plants adapted to very dry conditions tend to be polyploids. “The striking cytogeographic differentiation within the North American L. tridentata, together with the complete distinctness of the insect faunas on Larrea in North and South America,” wrote Phillip Wells and another scientist, “suggest that Larrea is of some antiquity on the North American continent.”
Axelrod’s botanist friend at UC Davis, Michael Barbour, thought creosote bush might have migrated from South America in a semiarid “transtropic scrub”: “Outlying pockets of Larrea in southern Mexico, in Chile, and in Peru intimate a historically wider range for Larrea. Minor bits of information sharpen that intimation.” Barbour observed that North American creosote bush seedling leaves tend to have three or more leaflets instead of two as with mature plants. Since South American nondesert species have three or more leaflets, this implied that our species’ ancestor might have spread north with “trans-tropic scrub” and then adapted to desert. “Stebbins (1952) believes that one of the consequences of adaptation to aridity for plants with compound leaves is a reduction in leaflet number,” wrote Barbour.
Axelrod disagreed:
The meager fossil record does not provide evidence for an arid corridor during the Tertiary. . . . Prior to the Pleistocene, arid areas were much smaller, less severe, and more widely separated than at present. Hence the concept of a “trans-tropic scrub” (Barbour, 1969), from which the present ranges of taxa such as Larrea may have developed, is unsupported by any evidence. The idea that relatively long ‘jumps’ are involved in the establishment of the North-South American desert disjuncts is in agreement with the observations that (1.) they constitute only a small proportion of their respective floras, (2.) the animals associated with them in their disjunct areas are almost entirely different and (3.) they are mainly self-compatible [able to pollinate themselves].
North American wood rat middens have yet to prove Darwin’s dispersal theory by disgorging a migratory bird mummy with South American creosote bush seeds in its crop. Given their fifty-thousand-year dating limit, middens don’t prove ideas of slower dispersal either. Without further evidence, California creosote bush origins are likely to remain obscure. But despite such limitations, and although their fossil preparation technique is not in the best of taste, wood rats have been surprisingly helpful scientific partners in revealing some of the desert’s past.