Well, I finished my Ph.D. in 1987. It got examined—that took forever. It was 1988 when it was listed. But when I tried to go from the primitive mantle, which is the tightest constraint, to the rest of the planet, I guess I got scared, if you will. I didn’t want to make any mistakes. Sometimes we have to live with our mistakes for a very long time in this business. So between 1987 and 1995, I had a considerable amount of trepidation about putting it all together as a statement and putting my name on it. In 1989, there was a paper by Sun and McDonough, and, in 1995, we had McDonough and Sun. The first, "Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes," (in Magmatism in the Ocean Basins, Spec. Publ. Vol. Geol. Soc. Lond. 42: 313-45, 1989. 2,448 cites, according to the Web of Science), was about basalts. You melt the Earth’s mantle, and you get basalts at the surface. We were trying to write a story about basalts. It wasn’t germane to that paper, but in that paper, and as a derivative out of the thesis, we published a statement about the abundance of about 25 elements in the primitive mantle. And we also said a McDonough and Sun paper would be forthcoming. That’s like saying a check is in the mail. It wasn’t truly misleading, but I hadn’t mustered it all together at that point. I was planning on it.

I probably had 20 papers or so.

It was kind of like keeping on task, but also taking on other tasks at the same time. In 1990, for instance, I published a paper on gallium. At the same time, I also published one on the composition of the continental lithospheric mantle. I needed to understand that better to get back to estimating the primitive mantle composition. So I went down a different road for a while just so I could understand my main highway better. In 1994, I wrote another one on isotopic systematics of the continental lithosphere. That was the same thing. In the interim, I kept refining my ability in this or that realm. In 1992, myself, Sun, and several big people in the business wrote a paper together on potassium, rubidium, and cesium in the earth and moon. Each step of the way I was increasing my confidence about my estimates, and I was being challenged by the community and I was challenging myself.

I can highlight two pivotal events. Around 1992, Ted Ringwood from Australian National University said he was really upset with me for having not published my paper yet. Ted passed away in 1993. He was an absolutely fantastic scientist. I kept telling him I was working at it but didn’t feel it was the right time. He said he was going to quote my Ph.D. dissertation, and asked if I minded him quoting my table. It was going to be an invited article in one of our premier journals. I was thrilled he wanted to do this. But publishing my entire table essentially publishes everything. His paper is cited in part for the numbers in that paper. Ted gave all due credit to me, but I knew after that I was committed to having to write and publish my paper pretty soon. Ted just referenced "McDonough Ph.D. thesis." About a year after that Kargel and Lewis published a paper in Icarus on the composition of the Earth, and did a mathematical model and cited my Ph.D. dissertation, albeit wrongly as a 1975 publication. Every element has a bin and you fill this bin up with whoever has published on that element. If you look at all their data, it was exceptionally dominated by my answers. It dominated their analysis. That said, I had no more time to wait—I’d better publish my paper.

You could say it provides a reference frame. A lot of people have models out there for a little portion of the Earth: say, rare earths in the upper mantle, or scandium content of the what-have-you. But people need a reference frame that covers everything, and so the potential exists for everyone to make this comparison between my paper and many other papers. Because my paper covers the composition of the Earth for all the elements in the periodic table and it compares it to the reference frame for what we believe to be a close approximation to the most primitive meteorites, mine provides the reference frame for all these comparisons. And then I hope the other reason they’re citing it is that the same table that has all the data also has uncertainty factors associated with each element. It tells you how well we know that element. So one reason they cite it so much is that they need a reference frame, and they can keep going to one reference frame that’s coherent in the sense that it discusses all the elements, not just some of them.

As for the second question, and this will sound horribly arrogant, I would say if we did this again today, we’d take the very same approach. That may be expressing a fairly pedestrian mindset on my part, but I can’t think of an alternative approach that will get you there. As for the first question, what’s happened since, there are a number of very large issues that remain unanswered. In fact, I would like one of my papers that came out last year to be almost as well-cited as the McDonough and Sun. In the last 10 years I have shifted a great portion of my academic research agenda to trying to refine compositional models of the Earth’s core. Last year I published a paper entitled "Compositional models of the Earth’s core" (in The Mantle and the Core [ed. R.W. Carlson] Volume 2: Treatise on Geochemistry [eds. H.D. Holland and K.K. Turekian], Elsevier-Pergamon, Oxford, 2003) trying to estimate that. It’s really difficult and I would say we are making progress, and we’re making progress in understanding the composition of the deep Earth’s mantle.

Well, for instance, there’s presently a lot of hubbub going on about what’s happening at the core-mantle boundary. That’s a very bizarre place. We’d like to know if there’s material transfer across that boundary. Is the metal of the core getting into the mantle, or is the rock of the mantle getting into the core? There’s a massive change in density across that boundary. It’s greater than that between the bottom of the ocean floor and the water above it. And through geophysical measurements and observations—mostly seismic wave measurements—we know that boundary is only a couple of kilometers thick. It may be even less than a kilometer. It appears to be very sharp on the bottom and diffuse on top. The reverse of what happens on the ocean floor. So one of the things that’s happened over the last couple of years is that some people have been saying that the material going from the core to the mantle is coming up in islands like Hawaii, just the tiniest amount of the Earth’s core coming out in Hawaiian lavas. And everyone has an opinion on that. I’m refining some of my models and weighing in on that. Other people are looking at their data and weighing in, and opinions differ, which is always the case in science.

I think it was to fix on an absolute value. This is going to be little complicated. There are something like 40 elements that are interdependent. So you fix one and you have the others fixed. Imagine that the whole solar system started out as a cloud between stars. Some supernova in the nearby neighborhood sends a shockwave through the system, and that gives this cloud a bit of spin momentum. So that imparts a center of mass, which becomes the sun and raises the temperature in this system. With time, you build planets. The first elements that start to condense out of that cloud are high-temperature elements, also called refractory elements, and that’s what these 40 elements are. The refractory elements are all in equal relative proportion in a wide variety of primitive meteorites. This is half of the periodic table. You fix one of these 40, you fix them all. So probably the hardest job was to actually go through my data and find out what was the best one or the best five to help me narrow down that fix. When I finally fixed it, it then said what all the other numbers had to be. Example elements would be calcium, aluminum, titanium, the rare earths, strontium, zirconium; it goes on and on. In the primitive mantle they’re all 1.16 times what you see in the most primitive meteorites relative to the magnesium content. That number, 1.16, I wanted it to be a number with very little uncertainty and with highly understandable traceable methodology. Once I had that number, everything else started to fall out from that. That was pivotal but that meant commitment also. I did that many times over, reinvestigated it, and it kept coming back to the same number. It kept hovering around 1.16, and that’s where it ended up.