The Reagan administration is considering development of weapons that could be transformed from conventional to nuclear with the insertion of a "clip-in" warhead, a technology rejected by Carter administration officials who thought it would undermine arms control and blur the distinction between conventional and nuclear war.

Almost any type of weapon can be developed to accommodate what the military calls "insertable nuclear components," according to designers at the Lawrence Livermore National Laboratory in California. The likeliest application would be to tactical weapons, such as Navy torpedoes or short-range Army missiles, according to the federal lab's arms control chief, Paul S. Brown.

Brown said development of such weapons would "create an arms control nightmare" because conventional and nuclear weapons would look alike. But he said "insertables," once again the subject of research at the weapons lab, might make nuclear arms easier to protect from theft or attack.

"I think it's going to drive arms controllers up the wall, but it does have obvious advantages in terms of security and safety," Brown said.

Richard N. Perle, assistant secretary of defense, said the clip-in warheads may make sense because the military could build a large number of missiles but only a small number of easily transportable nuclear warheads. As a result, fewer nuclear weapons might be needed, he said.

"If you could screw in the nuclear component to a Lance missile, it might be a really very promising way to economize in the number of nuclear weapons," Perle said. "And you get the same deterrent effect . . . . A very small number of nuclear components could go a long way," he added.

Perle disagreed that insertables would undermine arms control, saying that they would be most useful in small weapons that would not be covered by arms control agreements in any case. The administration has examined the application of clip-in warheads to both Navy and Army systems in recent years.

"There's nothing on the drawing boards at this point, but down the road it's something we should think about," Perle said. "The technology is now well in hand, and the neuralgia about nuclear weapons is such that there are more compelling reasons today than there were before."

Currently, some weapons such as cruise missiles are made in a nuclear and non-nuclear variant. But one cannot be converted into the other, and the two are not identical in appearance.

During the late 1970s, the Navy considered developing an insertable nuclear warhead for the Harpoon missile, a non-nuclear antiship weapon, according to Paul C. Warnke, then director of the Arms Control and Disarmament Agency (ACDA). The White House shelved the idea, partly in response to ACDA's objections.

"We decried the entire proposition, and I think we were successful," Warnke recalled this week. "Just in terms of having some handle on the nuclear threat, you'd lose it."

The technology was abandoned for a time, but designers at Livermore are reexamining it. The Army has experimented with a clip-in warhead for the Lance missile, and the Navy investigated a variety of uses for the technology with fiscal 1983 and 1984 research funds.

A Navy official said no insertable weapons are in production, but that the technology could be useful aboard ships and submarines, where space is at a premium. A submarine could carry 100 torpedoes and a footlocker with 100 clip-in warheads and not have to decide ahead of time how many conventional and how many nuclear weapons might be needed.

Richard Wagner, until recently the Defense Department's top adviser on nuclear weapons, said in a written response to a congressional query last year that clip-in warheads are a "militarily attractive option," particularly for the Navy. He said insertables would improve the "survivability and operational flexibility of U.S. theater nuclear weapons."

But William M. Arkin, a nuclear weapons expert and administration critic at the Institute for Policy Studies, said convertible weapons would make nuclear war more likely. Once conventional and nuclear weapons look alike, he said, an adversary may have to assume the worst.

"The other side sees Pershing missiles coming at him on his radar screen -- how does he know they only have conventional warheads?" Arkin asked. "It could have the impact of immediate escalation, because the other side won't know if they're nuclear-armed or not."

George H. Miller, director of the weapons program at Livermore, said research on convertible weapons is important in part to find out what the Soviet Union might be doing, "even if the United States decides not to pursue it."

"The fact that the Soviets could be doing it is important to know," he said.

Miller also said, however, that "there are lots of places it might make sense" for U.S. forces, including for strategic intercontinental missiles. A mobile missile like the proposed Midgetman could be driven around the highways, he said, if the warheads were held in reserve until a crisis.

"Warheads are easy to protect compared to a 30,000-pound or 100,000-pound missile," Miller said. "The arguments on the other side tend to be ones of countability," he said. "How many nuclear weapons have you got? It drives arms controllers crazy if they're bean counters."

Perle said the likeliest use for clip-in warheads would be on relatively short-range Army missiles.

"If they the Soviets face a front that is heavy with Lance missiles, and they can't tell if a Lance is nuclear or not, they will have to regard them all as nuclear," he said. "It's the only prudent thing to."

Such a capability, in turn, would force the Soviets to disperse their forces, Perle said.

The insertables would be small enough to be easily handled by troops in the field, Brown said, and Perle said they might be distributed only in a crisis. Arkin countered that the act of distribution might inflame a crisis.

Warnke said the government should be abolishing battlefield nuclear weapons, not creating new ones.

"What's the point of having this kind of capability, unless you feel that there's fungibility between conventional weapons and nuclear weapons and that we can fight and win a nuclear war?" Warnke said.

"What you're doing is building another one of these ambivalent systems that at least incrementally lowers the nuclear threshold," he said.