We compare the results of the observations and conclusions obtained from a remote science team and field geologists during the 1999 Marsokhod Silver Lake field experiment. This comparison is used to determine methods that enhance the ability of a remote team using a rover to determine the geology and paleoclimate of a landing site analogous to sites that may be encountered during the Mars Surveyor Program. The mobility of the field geologists and their ability to rapidly gain visual and physical data from the field site give an immediate confidence to their interpretation of the geology. The finite time of teleoperation and the initial decisions and mechanical problems that caused the rover to linger and collect data in a small area of the test site for the first 5 days of the test limited the ability of the remote science team to gain a similar confidence. Nevertheless, by appropriate use of orbital and descent images as well as the rover's science payload, the remote science team interpreted the present and former climates, the geology and geological history, and present-day biology relatively accurately. The few erroneous interpretations by the remote team resulted in part from biases brought to the test about the nature of the geology that was likely to be encountered. Data quality and variety also impacted the quality of the geological interpretations. Significant misinterpretations in the remotely determined geology resulted from the lack of detail in the optical imagery, a less than optimal understanding of the site's topography, and misinterpretation of some infrared data. All in all, this field test showed that the geology, paleoclimate, and biology of a site can be assessed to a reasonable level of confidence using a remotely controlled vehicle. It also showed that it is both feasible and desirable for a distributed science team to participate in long science missions from home institutions, as long as data flows efficiently to and from these institutions. ¿ 2001 American Geophysical Union