We have worked in closely related fields ever since, met at conferences all over the world more often than I can remember, visited each others' laboratories and worked together in connection with this journal. I hope this gives me the right to say something about her scientific work.
Della Roy came into cement chemistry from mineralogy. Her early contributions were in the field of hydrothermal chemistry, in which I also was active at the time. She quickly distinguished herself by synthesizing three new calcium silicate hydrate phases, all of which were later found as natural minerals. One of them, Dellaite, was named after her. This work, initiated for its mineralogical interest, proved to have important implications for the high-temperature uses of cements, including the production of autoclaved building materials and the cementing of deep oilwells and geothermal wells. From that time on, her ability and dedication in the field of cement chemistry were apparent and increasingly widely recognized.
Space permits me to mention only a few highlights of her subsequent work. With Mike Grutzeck, then a student under her direction, she made one of the first effective studies of cement paste using an electron microprobe, reported at the International Symposium held in Tokyo in 1968. During the 1970s, a series of excellent studies on very high-strength cement pastes, made by hot-pressing, provided a major stimulus to studies elsewhere on the factors governing the mechanical properties of cement-based materials. During this same period, in collaboration with Fed Tam,is, of Veszpr6m, Hungary, she greatly improved and extended a chemical method for determining the structures of silicate ions, applicable not only to crystalline materials but also to the nearly amorphous ones formed when cement reacts with water. The paper by Tam~s, Sarkar and Roy, published in 1976, is still widely quoted as describing the most and probably only reliable technique for this purpose. The results for cement pastes given by its use have been fully supported by later studies using nuclear magnetic resonance, a method that was not avilable at the time.