The fragments were liberated by ablating the central, anchoring portion of the centrosome. The fragments from the posterior centrosome flew out toward the cell cortex faster than did those from the anterior centrosome. They did not, however, fly toward a single focal point, which was a feature of some earlier models.
The simple velocity data did not distinguish between more motors or stronger motors. But after a look at the variance—the differences in the speeds of fragments from one experiment to the next—the “very messy data turned into beautiful, simple data,” says Howard. Variance initially increased, but at the highest velocities actually dropped. This pattern is seen in two-state (on and off) processes as they progress from frequent, noisy switching between on and off toward consistently on states.
If motors are only ever on or off, that makes further interpretation of the data possible. At a given mean velocity, the variance was higher at the posterior. This higher variance arises because there is a greater number of noisy elements—in other words, more motors—contributing to the final value. Which all goes to show, as Howard says, that “one person's noise is another person's signal.” ▪