Documenting the presence of magma: Robert Allen Dunn, postdoctoral research associate in geological sciences, and colleagues have discovered that there actually is molten rock in an area below the sea floor where some scientists thought there was none.
Dunn’s research team found evidence of the presence of magma along the East Pacific Rise, a mid-ocean ridge off the west coast of Mexico. The ridge began to attract attention in the 1970s when new life forms were found along the hydrothermal vents made possible by the heat.
The 60-mile-long ridges are interrupted by a break known as an overlapping spreading center. There’s been a longstanding controversy about whether or not there is magma coming up in the overlapping spreading center, and why these breaks form.
Dunn’s research provides the first images of the system of magma just below the crust. The research was published in the March 9 issue of Science magazine.
"The images, analogous to X-ray tomography or a CAT scan, provide visual maps of the thermal structure and distribution of magma in the Earth," Dunn said.
To create images of the magma, an airgun towed behind a ship sent a pressure wave through the water and into the ground. The seismic energy was recorded on instruments on the sea floor. Details are available online.
Dunn’s team presented preliminary results at the fall meeting of the American Geophysical Union in San Francisco. Dunn also presented research at the Northeastern Section of the Geological Society of America March 12-14 in Burlington, Vt. — Janet Kerlin
Building a better erythromycin: Chemistry Professor David Cane and colleagues at Stanford University have shown that the E.coli bacterium can be genetically manipulated to produce advanced chemical building blocks of the broad-spectrum antibiotic erythromycin. The antibiotic is widely used for respiratory infections.
Using these methods should someday allow the production of improved erythromycins. E.coli provides a much more powerful method for modifying the way erythromycin is made than the soil bacterium now used for commercial erythromycin production, Cane said.
"The need for new antibiotics has become increasingly urgent with the widespread occurrence of bacterial resistance to commonly-used antibiotics," Cane said.
The development of modified erythromycins is being pursued by Kosan Biosciences, where Cane is a member of the scientific advisory board as well as a shareholder.
The research was published in the March 2 issue of Science magazine. — Janet Kerlin
Word recognition - statistically speaking: Words for the favorite flower of Valentine's Day and the action of standing from a seated position are unrelated except for the sound sequence they share: "a-rose."
Programming a speech recognizer - a device that, when spoken into, produces the words on a computer screen - to determine whether the verb or the noun was intended requires more of a knowledge of statistics than that of grammar, said Mark Johnson, professor of cognitive and linguistic sciences.
Johnson discussed the shift in the study of language learning and language use to an optimization-based approach at the annual meeting of the American Association for the Advancement of Science (AAAS) in mid-February.
Optimization refers to the statistical frequency a word is used in language -how often it appears with a direct object and without a direct object, for example.
"It's a big change in perspective to view language as a process of optimally identifying characteristics," said Johnson.
By looking at the statistical properties of words, researchers can analyze large bodies of text, such as all the articles published in the Wall Street Journal over five years, and form conclusions about the frequency with which individual words are used in language.
That type of study requires collaboration between researchers in the departments of applied mathematics, computer science, and cognitive and linguistic sciences. University support for interdisciplinary research at Brown has facilitated the study of computational linguistics here, said Johnson.
The implications for language optimization research range from speech recognizers to Internet search engines, which rely on key words to locate documents, said Johnson.
So is a-r-o-s-e still a rose? Depends on the words around it, said Johnson.
"How a computer would know whether or not to put a space between a- and r-o-s-e depends on a statistical model," said Johnson. "Probability is a factor. Where is the two-word sequence more likely?" — Kristen Cole