Dinosaurs have certainly become a lot prettier in the decades since I was a child. But I wonder if these splashes of color, such as seen in the images in “Triumph of the Titans,” by Kristina A. Curry Rogers and Michael D. D'Emic, have gone too far. Among living creatures today, none of the larger land animals sport elaborate color schemes. And in the smaller critters, color has evolved as a camouflage, something that would have little or no success on an animal the size of an elephant—or an Apatosaurus!

John Byrne
via e-mail

THE AUTHORS REPLY: We chose to “spruce up” our depictions of sauropods with some color to reflect our new understanding of these animals as fast-growing and dynamic, with a unique mix of birdlike and reptilelike biology. Although most large-bodied animals today could be construed as drab, some are indeed colorful (giraffes and orcas come to mind). As for camouflage, the bright colors of many small animals have not evolved for that function but for the purposes of attracting a mate or signaling kin or a predator.

Notwithstanding, it is important to remember that sauropods did not start off life large (baby sauropods were about half a meter in length), so predation was a factor in a young sauropod's life. Ultimately the coloration of most dinosaurs remains speculative, although recent research has indicated complex patterns and colors on some feathered dinosaurs.


In “The Trouble with Wi-Fi,” by David Pogue [TechnoFiles], Don Millman of Point of Presence Technologies responds that the reason high-end hotels charge for Wi-Fi is that they “attract business travelers who expense their stays.” While this is not incorrect, the answer is more complex. For example, most high-end resorts serve more leisure travelers than business guests but charge for Wi-Fi. High-end hotels are mostly operated by brand management companies with fees based on a percentage of total revenue; most less expensive hotels are franchised, and their fees are based on room revenue only. Thus, there is an incentive for brand management companies to maximize their revenue—for example, by charging for Wi-Fi—and an incentive for less expensive franchised hotels to have higher room rates—for example, by including Wi-Fi in the rate.

But the most correct answer is that guests in more expensive hotels are more willing to accept charges for Wi-Fi. A reason to charge separately for Wi-Fi is that if it were included in the room rate, municipal occupancy taxes would be based on the higher room rate, and therefore the rate would be more costly for guests.

Bjorn Hanson
Dean, Tisch Center for Hospitality,
Tourism, and Sports Management
New York University


The new unitarity approach, a method of analyzing quantum-particle processes less complex than the Feynman diagrams that have been the standard, proposed by Zvi Bern, Lance J. Dixon and David A. Kosower in “Loops, Trees and the Search for New Physics,” doesn't seem new. Applying estimates of probabilities to beginning and prior events in a causal chain to determine the final outcome's probability (adjusting as new results are available) simply describes Bayes' rule, doesn't it?

Duncan Byers
Norfolk, Va.

THE AUTHORS REPLY: If the quantities we were interested in were really probabilities, then one could assemble them as Byers suggests and it would not be particularly novel. As we mentioned in the article, they are really square roots of probabilities. These are complex numbers, although for simplicity we usually referred to them as probabilities. One combines them according to the usual rules of quantum mechanics, wherein the phases associated with them are essential. They capture quantum-interference phenomena that prevent the application of the usual probability rules. Another obstruction to simply multiplying probabilities in Feynman diagrams is the presence of the spurious contributions we had described in the article, which disallow a simple probabilistic interpretation of an individual diagram.

There are useful approximations where the interference terms can be neglected. These have been implemented as products of sequential probabilities in computer programs, widely used by particle physics experimenters, that produce realistic-looking jets of particles. But their overall accuracy is not as good as the loop calculations described in the article. Combining the best features of both techniques is a very active area of current research.


The lack of low-altitude radar coverage as a factor in tornado-warning lead times was not mentioned in “A Better Eye on the Storm,” by Jane Lubchenco and Jack Hayes. Even with improvements in signal processing and phased-array radars, the unobservability of the zone where tornadoes form will limit the ability to forecast. Not all urban areas have or will have radars close enough to observe the bottom 5,000 feet of highly active weather systems when tornadoes pose the greatest threat. Very small short-range radars might be co-located with cell-phone towers to provide the density, power, communications and required altitude.

Doc Dougherty
Playa del Rey, Calif.

The authors speak of new, higher-resolution weather models that use horizontal grids. Wouldn't it be better to concentrate computing power where it would be most valuable? Model designers might take a cue from image-compression software that uses fine resolution for only those areas in an image where changes occur. It should result in horizontal and vertical grids that fluctuate in fineness in response to actual and predicted conditions—lowering the number of data points in areas where weather is fairly uniform and increasing them where conditions are changing rapidly.

Ralph McLain
Colorado Springs, Colo.


The poplar and sugar maple experiment by Ian Baldwin and Jack Schultz, described in Daniel Chamovitz's “What a Plant Smells,” doesn't actually demonstrate signaling between plants. As told, undamaged leaves on trees having two damaged leaves were the ones to respond to that damage by making caterpillar deterrents. Nothing is said of the behavior of the intact trees of the same population.

Dov Elyada
Haifa, Israel

CHAMOVITZ REPLIES: The confusion comes from editing that shortened the excerpt from chapter 2 of my book, What a Plant Knows. Indeed, Baldwin and Schultz detected insecticidal chemicals not only in the intact leaves of the trees that had torn leaves but also in the intact leaves of trees that neighbored them. This was the basis of their volatile communication hypothesis.


“Erasing Painful Memories,” by Jerry Adler, incorrectly refers to a foot shock causing a rat to avoid a particular area as a negative reinforcement. It is more accurately described as a positive punishment.