Populations of salamanders and aquatic insects, animals sensitive to pollution, were 25 percent lower in streams with the most drilling activity.
An industry spokesman declined to comment on the findings.
The researchers at the academy, the nation's oldest natural-science research center and a leader in stream biology, emphasized that their study was not looking at drilling accidents or other irregularities, but whether - and if so, at what point - drilling posed a potential for harm.
David Velinsky, vice president of the academy's Patrick Center for Environmental Research, said of the early findings: "This suggests there is indeed a threshold at which drilling - regardless of how it is practiced - will have a significant impact on an ecosystem."
A certain number of well pads in a given area "might be OK," he said. "Conversely, it may not be OK."
The intent of the research, he said, is "to try to find where that stands, to look at cumulative impacts across a gradient of drilling."
The academy researchers are using the study, which has not been peer-reviewed or published in a scientific journal, as the basis for requesting state funds to do a much larger and more comprehensive study, Velinsky said.
Travis Windle, a spokesman for the Canonsburg, Pa.-based industry group Marcellus Shale Coalition, said in an e-mail that the group "does not comment on preliminary, non-peer reviewed, unreleased 'studies' that we have not even had the opportunity to examine."
He did say that total dissolved solids - another often-used indicator of salt contamination - in a stream was not necessarily the fault of drilling.
"Pennsylvania has a long history of high TDS levels in our waterways - long before Marcellus production commenced just a few years ago," he said.
Velinsky agreed that there are other sources of salts and solids, including road salt and fertilizer.
But "we tried to control for other inputs of salt as best as possible in this limited study. We saw this relationship" between drilling and stream changes, he added, "and it warrants further study."
The Patrick Center, formed by the world-renowned ecologist and limnologist Ruth Patrick, is staffed by scientists and engineers. It focuses on understanding aquatic and terrestrial ecosystems and using the information to develop strategies to enhance watersheds.
Drilling for natural gas in the Marcellus Shale formation has become a controversial issue and a political hot button across the state.
It is viewed as an economic boon and a potential solution to the nation's dependence on foreign oil. Since natural gas burns with fewer emissions than other fossil fuels, it also is considered an answer to climate change.
However, critics cite the potential for environmental damage, the industrialization of urban areas, and the loss of a natural resource - owned, in some cases, by the public - to for-profit companies.
The legislature is grappling with the issue of whether to levy a state tax on natural gas extracted from the Marcellus Shale.
While rhetoric on drilling has been abundant, Velinsky said that independent scientific studies on the environmental effects have been scarce: "There are a lot of voices out there that are going from 'Oh, it's no problem' to the other side, 'It's doom and gloom.' "
"We wanted to get some good, hard science information, some good data, that could help eventually guide the managers - either the gas companies or the Pennsylvania Department of Environmental Protection or the various other regulators - into what may be a good process for the drilling," Velinsky said.
Velinsky said he knew of no other similar studies that have been completed.
The Heinz Foundation is funding a three-year, $2 million study to create a baseline survey of aquatic and terrestrial ecosystems that will be affected by drilling in the Marcellus Shale.
It involves researchers from Duquesne University, the University of Pittsburgh, Carnegie Mellon University, California University of Pennsylvania, and the Carnegie Museum of Natural History.
As part of that, a team will study the salamanders, fish, and microbes in Washington County's Ten Mile Creek watershed in southwestern Pennsylvania.
As of Oct. 1, the state had issued 5,095 permits for Marcellus Shale gas wells, according to DEP records, which also show that 2,237 Marcellus wells have been drilled.
"We recognize that drilling is going to happen," Velinsky said. "There are a lot of pluses for it to happen. . . . But, hopefully, it can be done in an environmentally sensitive way."
In the academy's preliminary study, a University of Pennsylvania graduate student, Frank Anderson, studied nine small watersheds in northeastern Pennsylvania.
All were similar in terms of land use and the size of the streams. But they differed in one primary area - the amount of natural-gas drilling.
Three watersheds had no drilling.
Three others had what the researchers described as a low density of drilling, which is one to three wells per square kilometer of land area.
The other three had a high density of drilling, or four to eight wells per square kilometer.
A square kilometer is slightly more than a third of the area of a square mile.
Anderson then measured the conductivity of the water and assessed the abundance of salamanders, mayflies, caddis flies, and stone flies.
Now, the researchers want to conduct a two-year study and expand the sampling from nine streams to 36, said Jerry Mead, a systems ecologist with the Patrick Center who worked closely with Anderson.
"If drilling is restricted to a certain density," Mead said, "maybe we can have some drilling and get the benefits for society and not have a huge impact on the stream system and all the services it provides people."
Contact staff writer Sandy Bauers at 215-854-5147 or email@example.com. Visit her blog at http://go.philly.com/greenspace.