The Earth's orbit around the Sun isn't perfectly circular. And, like a spinning top, the Earth has wobbles. The overall result is that there are a number of cycles, and these cycles should affect the amount of sunlight we get.
Astronomers long ago worked out that the most detectable cycles would take 23,000 years, 41,000 years, 100,000 years, and 404,000 years. So, if we find a layered rock, it would be interesting to measure the layer thicknesses, and study the numbers to see if they're cyclic. (That's called a Fourier analysis.)
Specifically, it would be very significant if one found a cycle of 23,000 layers. And it would be just 'way beyond coincidence if there were three cycles, and they were in the ratio 23:41:100. It would be even better if the cycles repeated many times. Exactly such evidence is reported in
Clemens & Tiedemann
Broecker & Denton
except that they didn't measure thicknesses. In fact, they drilled a long sample - a core - out of the ocean floor. At a great many places in the core, they measured an isotope ratio. The ratio changed in just the cyclic way we've been talking about. So, just on the astronomical evidence alone, we can say that Clemens' sea-floor sediments took 4,000,000 years to form, and Broecker's sea-floor sediments took 600,000 years to form. The samples were also dated, and the shortest cycle was dated as taking 23,000 years. This is a very nice confirmation that radioactive dating works well.
Those two ocean sediments aren't the only cases:
In recent months a 25 million year long record from the triassic (about 200 million years ago, for those of us who believe such things) has been obtained. The rock is banded, and the bands form quite regular groupings. The smallest bands contain about 20,000 varves (annual layers) - and the precession cycle at that time was about 20,000 years long. Coincidence? Well, the precession cycle is modulated by the 100,000 year eccentricity cycle so the bands should occur in groups of five, with slightly different characteristics within the group. They do. Not enough? There is also a 400,000 year eccentricity cycle, so the large bands should be bunched in groups of four. And they are.The Scientific American article mentions Dr. Hyde's work on exactly how the astronomical cycles affect ice sheets.
Astronomy, not radiometric dating, tells us that this sample of rock was laid down over 25,000,000 years. So the earth is at least that old. Furthermore, since K-Ar dating gives the same length to this record we have no reason for not trusting within a few percent the K-Ar absolute age for this stratum, which is about 200 million years.
The geological evidence was presented by Paul Olsen of Lamont-Doherty at a recent workshop at Johns Hopkins. The theoretical paper isShort, D. A., J. G. Mengel, T. J. Crowley, W. T. Hyde and G. R. North 1991: Filtering of Milankovitch Cycles by Earth's Geography. Quaternary Research. 35, 157--173.
-- Bill Hyde, Department of Oceanography, Dalhousie University, Halifax, Nova Scotia, 19 Mar 1994