The principles we've covered so far are basically what were used to determine the order of the rock strata (particularly superposition and faunal succession/biostratigraphy). This wasn't simple or easy and it certainly didn't happen overnight, but with thousands of observations at thousands of localities a picture was built up of the extents of various formations and their relationships to one another.
So how do we test if we got the order correct?
Radiometric dating.Radioactive isotopes are generally unstable and over time decay from their 'parent element' to a more stable form, the 'daughter element'. We can measure the rate at which this decay happens* and work out what is known as the half life. This is the time it takes one half of the parent elements to decay, and it forms an exponential curve. After one half life is over, half of the parent element remains. After two, a quarter, three an eighth, four a sixteenth etc etc. We determine the age by measuring the ratio of parent to daughter elements and see how many half lives of known length this equates to. There are a number of different elements we can use for dating, many with different half lives. Those with shorter half lives are suitable for younger material, and those with longer half lives for older. Misusing them will give erroneous results, for example carbon14 is good up to about 70,000 years. Try to date palaeozoic coal with it and it will give you a younger date. Likewise potassium argon dating is only good on material over 100,000 yrs. Try to date lava that is practically still warm and again you'll get an incorrect result.
We can typically only date igneous material with most of these methods, however carbon dating is for biological material. Some metamorphic material may be dated with some techiques, but this will give the time of metamorphosis, not when the parent rock (rock that was metamorphosed) was formed.
Here is a very good resouce on radiometric dating which deals with a lot of misconceptions about it as well as explaining numerous techniques.
I don't want to dwell on this too much, but the point I'd like to make is that radiometric dating is an independent method that confirms the ordering of rocks as determined by other methods.
For some more evidence that this dating is valid, I'd like to borrow from
this excellent thread by Loudmouth.
Loudmouth writes:
The Hawaiian islands and the Emporer seamounts are the direct cause of volcanic activity. It has been theorized that these formations all originated from one mantle plume. This same mantle plume is now responsible for the eruptions at Kilauea. Also, the Pacific plate, on which the seamounts and islands are part of, is moving northwest at about 8-10 cm/year as measured by GPS. So, as the Pacific plate moves over the hotspot, new islands will crop up in a line. The picture below shows the Hawaiian chain and the Emporer chain, with Kilauea being the last island on the right hand side:
Conventional geology, through the theory of tectonics, puts forth the following predictions:
1. The measured age of the seamounts will increase, in a linear fashion, as the distance from Kilauea increases.
(snip)
YEC theories, saying that these seamounts and islands were created within a short geologic time frame, put foth the following predictions:
1. There should be no linear correlation between the distance from Kilauea and the age of the islands or seamounts. This is because YEC's believe that such islands were made quickly, and not over millions of years. Also, they believe that K/Ar dating is not accurate. Therefore, no realtionship should be found between a faulty dating methodology and real life islands.
(snip)
AGE:
The age of the islands was measured using K/Ar dating. Now, if either the plate movement, radioactive decay rates, or the precision of the dating method were not reliable, then we should not see a linear relationship between the measured age of the islands/seamounts and their distance from the hotspot (ie Kilauea). However, there is a linear relationship as seen in this graph:
You can find the actual numbers for the graph
here if you want to construct your own graph.
Going with the data, we can see that the Pacific plate moved about 5,000 km over a 65 million year span. If at any time the decay rates changed, then we should not see a linear relationship. Instead, there should be an abrupt change in the age/distance ratio. Also, if there was a great shift in the plate movement, this relationship also should not be linear.
(edited out material not relevant to this thread where (snip) is written)
It seems apparent that radiometric dating has confirmed that we have the order of rocks correct.
*Just to clarify, it's not possible to predict when an individual atom will decay (correct me if I'm wrong here someone), however in their millions they behave predictably. Picture unweighted dice. If I were to roll and told you to guess the outcome, it would be purely by luck if you were correct as its is essentially random, however if I were to roll it a thousand times, you could bet good money that any given number would be rolled approximately one sixth of the time.