He and his colleagues climbed to 9,000 feet, where they entered chill, foggy Oyamel fir forest. "We walked in along this trail that went down along a ridge for a mile, and we didn't know exactly where the butterflies were," Brower recalls. Then, abruptly, the deep greens of the firs were lost in a dazzling cataract of orange, and Brower realized he was looking at a wall of butterflies. "I couldn't believe the density and numbers," he says. "The next day it warmed up, and when the sun came up, the butterflies were flying all over the place. It was like walking into Chartres Cathedral and seeing light coming through stained-glass windows. This was the eighth wonder of the world."

It also was proof positive that this was indeed the wintering ground of migratory monarchs. These orange and black insects, which live in the tropics of Central America and South America as well as in temperate North America, are divided into two populations in the United States. Those west of the Rockies winter along the California coast from just north of San Francisco to just south of Los Angeles. During spring and summer, they migrate across California's Central Valley and breed in the Sierras and northward to Oregon.

Monarchs east of the Rockies migrate between Mexico and the Great Lakes, swinging 2,000 miles across much of the eastern United States and southern Canada. But despite the great distances involved, during the past 20 years this migration has become better understood than that of western monarchs, thanks to persistent research by scientists such as Brower.

Much of what biologists know is brand new. Not until just two or three years ago was Brower able to provide insight into how monarchs navigate during their travels and just where they spend their time above the U.S.-Mexico border. One of Brower's hypotheses--that monarch butterflies plot their route with an internal magnetic compass that shifts a full 360 degrees each year, yielding a great migratory circle--is so new that Brower is still gathering data to see if it is correct.

Lepidopterists have been watching throngs of fluttering monarchs since at least the 1850s, but not until 1930 were scientists able to conclude with certainty that monarchs migrate south for winter and north in early spring. Precisely where they went for the winter remained a mystery, a kind of entomological Holy Grail.

In 1940, Toronto entomologist Fred Urquart searched for the winter home by setting up a network of more than 3,000 butterfly lovers scattered across North America. The group tagged monarchs in order to map their movements over the course of each year. As time passed, Urquart received reports of tagged butterflies turning up farther and farther south, first in Texas and then in northern Mexico. His searches for their wintering sites, however, always ended in frustration.

Then, in 1973, Kenneth Brugger, an American businessman living in Mexico City, read of Urquart's work and wrote to him about seeing monarchs pelted out of the sky during a hail storm in the mountains west of Mexico City. Urquart urged Brugger to explore the region. Brugger, with his Mexican wife, did so, and on January 2, 1975, they found the winter home of the monarchs, high in the Transverse Volcanic Belt. In this 20-by-40-mile strip of mountains, hundreds of millions of monarch butterflies find refuge in a dozen patches of forest at elevations of around 10,000 feet. There they congregate from November to March on trees in densities as high as 4 million butterflies per acre.

Research by Brower and other biologists indicates that these fir forests uniquely meet the monarchs' special requirements for a winter home: a place cool enough to put the butterflies' reproductive urges on hold until the next breeding season, yet not so cold that the insects freeze. Above 7,000 feet, night-time temperatures in the Transverse Volcanic Belt drop to as low as 25 degrees F. On open ground, the butterflies would die from such cold, but in the mountains they cling to the firs, shielded by the canopy from rain and snow and warmed by heat rising from the ground at night.

In addition to studying monarch winter ecology, Brower has perfected a technique for tracking monarchs throughout the rest of the year, based on the butterflies' special relationship with milkweed plants. Monarchs lay their eggs on milkweed and, as larvae, eat the plants not only for nutrition but also for the toxins in the sap, which is poisonous to birds that prey on butterflies. The toxic compounds, known as cardiac glycosides, accumulate in larval tissues and persist after the larvae metamorphose into butterflies, making monarchs unpalatable to birds. "I've tasted the sap of the sandhill milkweed," says Brower, "and it's incredibly nauseating."

Brower realized in the mid-1970s that he could use these poisons to understand monarch migration. The specific glycosides present in different species of milkweed vary from place to place, so analysis of a plant's glycosides can be used to determine its site of origin. Likewise, the glycosides in adult butterflies are a fingerprint of the plants on which the insects fed as larvae. "The light bulb went on," says Brower. His idea: If monarchs reflect the different arrays of glycosides in milkweeds, then he should be able to analyze butterfly glycosides and correlate individual insects with particular species of milkweed and with certain regions where the milkweed grew.

Traveling around the country, Brower trapped monarchs everywhere from the Great Lakes to Mexico and studied their glycosides. "We collected monarchs in the fall migration and established that over 90 percent fed on the northern milkweed pattern," he says. "Then we sampled the butterflies in Mexico and the ones along the Gulf Coast at the end of March and early April and, lo and behold, all of them had the same pattern. The same butterflies were going to Mexico and coming back in the spring."

Brower's work on glycosides helped to reveal the entire pattern of monarch migration. The first generation of monarch butterflies to return from Mexico stops along the Gulf Coast from Texas to Florida, where the insects lay eggs and die. The eggs hatch, and the new monarchs (with their distinctive southern milkweed fingerprint) head for the Great Lakes region. Later, the new generation of monarchs born around the Great Lakes flies to the East Coast, where they breed and die. Their offspring, fed on eastern milkweeds, then head south toward the Gulf Coast and on to Mexico, traveling as far as 90 miles per day and completing the great circle of monarch migration.

What used to be a confusing rampage of butterflies now appears to Brower to be a stately circle around the eastern United States. "The natural history of these things suggests that they're migrating throughout all their generations and just changing the angle of their orientation," Brower says.

That elegantly simple flight path might explain how the monarchs manage such an incredible journey. Many birds use the Earth's magnetic field to orient during migration, and some species change their orientation to the field seasonally. In spring, the birds' internal compass is set for north, in winter, for south.

Brower's research suggests that monarch butterflies spin their compass harder than does any other animal. When they go to Texas in spring, their flight is oriented due north. But the next generation heads northeast to the Great Lakes, and their offspring go east over the Appalachians. The following generation has its compass pointed southeast initially but, during the long journey, appears to turn south and then west, bringing the monarchs to Mexico. In other words, successive generations of monarchs spin the migratory compass in a complete circle over the course of a year. Brower suspects that changing day length cues butterfly orientation, and he soon will conduct experiments to see if he is right.

Going into his fifth decade of research on monarchs and supported now by the Wildlife Conservation Society, Brower recognizes some very practical reasons for knowing exactly where the butterflies go and how they get there. The vast prairies that once were full of milkweed have been plowed under for many decades, and now U.S. farmers and road crews use 620 million pounds of herbicide yearly to stifle the growth of native plants. "If you go out to the rich agricultural parts of the United States, the use of herbicides is so extensive, even along dirt roads, that they completely kill all wildflowers, including milkweeds," Brower says.

At the same time, the Mexican mountain forests that protect the monarchs each winter, located in a region where 500,000 poor farmers struggle to survive, are being nibbled away by logging. Even five sites protected by presidential decree, in which logging is illegal, are put to the axe, though the Mexican goverment strives to enforce its restrictions on tree cutting. The insects are vulnerable to logging since even thinning lets out enough heat and lets in enough rain and snow to put the butterflies in jeopardy.

Although scientists have not yet detected a steady decline in monarch butterfly numbers, the threat of habitat loss is so clear that last August the United States signed an agreement with Canada and Mexico to begin looking into the designation of reserves where monarchs will be sure to find intact fir forests and fields of milkweed. Canada has already set aside three sites in southern Ontario through which monarchs regularly pass.

Although the agreement is an encouraging sign, many years may pass before it affords the monarchs real protection. Yet only if these remarkable creatures have safe passage on their year-round journey will their vast migratory circle remain unbroken.

Bill Rankin, a frequent contributor to National Wildlife, works as a magazine editor and freelance writer in New York City.