News of the Week majors). After having risen sharply in 1987, demand for B.S. chemists (by the reporting companies) will decline about 11% in 1988. Starting salaries for B.S. chemists will average $25,692, an 8.8% gain from 1987. At the M.S. level, reporting firms7 hiring plans call for a 10% increase from 1987. Demand for M.S. engineers will be up 6% from 1987, and their average starting salaries will rise 3.3% to $34,776. Demand for nonengineering M.S. graduates will increase 14%. The highest-paid group among new masters will be M.B.A.s with a technical B.S., with average salaries rising 10.4% to $38,412. Ward Worthy, Chicago
Carbide ordered to pay Bhopal victims The presiding judge in the litigation against Union Carbide in Bhopal, India, has ordered the company to donate $270 million as an interim sum for the relief of victims of the December 1984 chemical disaster. Carbide reacted quickly but tentatively to the decision by Judge M. W. Deo in the central Indian city last week. "I haven't seen the text of the order/' said Bud G. Holman, attorney for Carbide. "But we've long been of the view that there is no basis whatever in the law for any award of interim compensation/ 7 He said such judgments must be based on evidence presented through trial procedures. Holman also said that Carbide already had made offers of interim relief, as opposed to outright compensation, and that local measures to help relieve the plight of the gas victims had already been deemed adequate by Indian authorities. So, drawing a legal line between relief and compensation, he said relief did not seem to be the issue. Still, he said, "If there are some aspects to Judge Deo's order that would facilitate a fair and responsible settlement, then that's to the good. The judge back in early April suggested a substantial interim relief amount and that was a catalyst for settlement." But Holman added that Carbide would resort to legal means to coun6
December 21, 1987 C&EN
termand Deo's order if it amounts to compensation in the legal sense. That would mean appeal to the supreme court of the state of Madhya Pradesh, where Bhopal is located, and, probably, recourse to American federal courts on the basis of failure to receive due process. According to wire service dispatches, Deo said his order was not meant to prejudge Union Carbide, the implication being that if Carbide were to convince the court of its innocence, the money would be refunded. Carbide wants no implication that it is blameworthy. Earlier this month, the Indian government filed criminal charges against Carbide and its Indian affiliate just as settlement talks broke down. India wants a settlement package amounting to $1 billion. But politically powerful protest groups are insisting that Carbide also be taken to trial to determine guilt. Wil Lepkozvski, Washington
NMR screening test for cancer questioned A screening test that has been reported to detect cancer in patients through nuclear magnetic resonance studies of their blood plasma is being viewed with growing skepticism. The report made newspaper headlines last week, but scientists from several research groups are saying they have been unable to confirm the NMR experiments. The work in question was first published late last year by Eric T. Fossel, director of radiology research at Boston's Beth Israel Hospital, and his coworkers [N. Engl. J. Med., 315, 1396 (1986)]. Earlier this month, Fossel presented the results of an expanded study of more than 3000 individuals to a symposium cosponsored by the French Association for Cancer Research and the National Institutes of Health. By suppressing the strong signal from water that usually dominates NMR spectra of blood plasma, the researchers are able to examine the methyl and methylene resonances of lipoprotein lipids. They claim that the linewidths of those peaks are significantly narrower in plasma
from patients with malignant tumors than from healthy subjects or people suffering from other diseases. Other scientists, however, do not see such a clear-cut distinction. "If it worked, it would be marvelous," says Peter Wilding, head of clinical chemistry at the University of Pennsylvania medical school. Wilding has a paper in press in Clinical Chemistry that details his group's unsuccessful attempts to replicate Fossel's work. "We do see a difference in average linewidth between normal individuals and patients with malignancies," he says. "But there is so much overlap that there is no clinical utility." In addition, Wilding finds that plasma from people with high blood levels of triglycerides also shows narrow average linewidths. "Using Fossel's criterion, those patients would have been diagnosed as having had malignancies," he says. Scientists at Philadelphia's Fox Chase Cancer Center have encountered the same problems with overlapping linewidths. "Here we've been totally unable to reproduce Fossel's results," says Truman R. Brown, head of the NMR department. "The technique is utterly useless. I'm worried that it will suck up valuable resources that ought to be going elsewhere." Fossel suggests that the discrepancies may be due to the other researchers' inability to duplicate his experimental conditions. The temperature, sample collection procedure, and magnetic field profile are crucial, he says, and several labs have confirmed his work. His critics, however, remain unconvinced. For instance, Jack S. Cohen, head of the National Cancer Institute's section on biophysical pharmacology, says he and a colleague have decided not to pursue an NMR study they began considering when they first heard of Fossel's work. "From what I've learned from others who have tried to duplicate this NMR test, I am not prepared to commit my time and resources to a study of it," Cohen says. "Some people believe this is real and therefore interpret their results that way. It's a general problem in science of people seeing what they want to see." Pamela Zurer, Washington