According to Freud and Kinsey, some people are sexually attracted to both men and women.
A new study casts doubt on whether true bisexuality exists, at least in men.
This new study, by a team of psychologists in Chicago and Toronto, backs those who have long been dubious that bisexuality is a distinct and stable sexual orientation.
Critics say that men who claim to be bisexual are usually homosexual, but are either ambivalent about their homosexuality or simply still in denial, and many gay men will insist a person is either gay, straight or lying.
In this study, a team of psychologists were quite direct in their methods and measured genital arousal patterns in response to images of men and women.
The psychologists apparently found that men who identified themselves as bisexual were in fact exclusively aroused by either one sex or the other, and that was usually other men.
This particular study is the largest of several small reports which suggest that the estimated 1.7 percent of men who readily identify themselves as bisexual, in fact show physical attraction patterns that appear substantially different from their professed desires.
Dr. Lisa Diamond, an associate professor of psychology and gender identity at the University of Utah, who was not involved in the study, says research on sexual orientation has in the past been based almost entirely on self-reports, and this study is one of the few good studies using physiological measures.
Diamond says that the discrepancy between what is happening in people's minds and what is going on in their bodies, is puzzling and it raises the question what do people mean when they talk about desire.
She says this is evidence that the assumption that everyone means the same thing is not so.
Other researchers however say it would need to be repeated with larger numbers of bisexual men before clear conclusions could be drawn.
Dr. Randall Sell, an assistant professor of clinical socio-medical sciences at Columbia University, says that bisexual desires are sometimes transient and are still poorly understood.
Men and women also appear to differ in the frequency of bisexual attractions.
He is concerned that some therapists might use the study's findings as reason to tell bisexual people that they are really on their way to homosexuality.
He feels too little is known about sexual orientation and identity, to jump to these conclusions.
For the experiment, psychologists at Northwestern University and the Center for Addiction and Mental Health in Toronto used advertisements in gay and alternative newspapers to recruit 101 young adult men.
Thirty-three of the men identified themselves as bisexual, 30 as straight and 38 as homosexual.
The men were questioned about their sexual desires and rated them on a scale from 0 to 6 on sexual orientation, with 0 to 1 indicating heterosexuality, and 5 to 6 indicating homosexuality.
Bisexuality was measured by scores in the middle range.
The men were seated alone in a laboratory room and watched a series of erotic movies, some involving only women, others involving only men.
The researchers by using a sensor to monitor sexual arousal, found, in line with their expectations, that gay men showed arousal to images of men and little arousal to images of women, and heterosexual men showed arousal to women but not to men.
However the men in the study who described themselves as bisexual did not have patterns of arousal that were consistent with their stated attraction to men and to women.
Instead, about three-quarters of the group had arousal patterns identical to those of gay men, while the rest were indistinguishable from heterosexuals.
Gerulf Rieger, a graduate psychology student at Northwestern and the study's lead author, says that regardless of whether the men were gay, straight or bisexual, they showed about four times more arousal to one sex or the other.
Rieger says that despite the fact that a third of the men in each group showed no significant arousal watching the movies, their lack of response did not change the overall findings.
The study is published in the journal Psychological Science.