Fog-laden Patrick's Point juts into the Pacific Ocean, home to centenarian redwoods presiding over the marine campground and hiking paths. One path leads to a small cove beach, protected by cliffs. The sandless Moonstone Beach is host to countless pilgrims, young and old, who squat to comb the millions of rocks for moonstone - pure, white, unblemished, smooth. Their searches are mostly in vain; there is not a piece of moonstone in sight. The beach, however, is littered with jade. Had it been designated "Jade Beach," I would not have seen a piece of jade in sight. Visitors were not looking for any beautiful sea rock; they were looking for moonstone, and thus missed the cornucopia of other treasures under their feet.

Anytime we prescribe an agenda, a goal, a desired outcome for our inquiries, we blind ourselves. Looking for one result, we overlook another. Complete objectivity is impossible, simply because even the most rational people approach their investigations with a set of questions, an agenda. But isn't this how we define science - by seeking answers to particular questions? Without questions, what "science" is left? It is impossible, given the infinite mysteries our universe holds, to approach science without a guided destination. You have to test something. Science gets itself in trouble, however, when the agenda becomes more important than the data.

Science is facing an objectivity crisis. Digital media technology permits doctoring of images; and scientists have bitten into the forbidden fruit. Dr. Hany Farid, an associate professor of computer science at Dartmouth College, specializes in the field of detecting image doctoring. He laments that "[i]t used to be that you had a photograph, and that was the end of it - that was truth," but that today, amidst the abundance of digital media, images need to be subjected to the same scrutiny as the written word. How does the scientific community establish a structure for evaluating objectivity and validity of its publications?

Science has attempted to devise procedures - the Scientific Method, for example - to ensure that questions in scientific research never become more important than the findings they produce. The Scientific Method prescribes doses of observation and description, formulation of hypothesis, prediction of explanation based on hypothesis, and performance of experimental tests to verify the hypothesis. It is science's rational, emotionally detached approach to investigation that prompts our society to hold science as the model of objectivity. The scientific method is designed to systematically and objectively evaluate the accuracy of an observed phenomenon or explanation. In an ideal, theoretical, emotionless world, this might work.

Many scientists, however, have agendas. Scientists hypothesize, conjecture, anticipate, and hope; when results don't fit speculations, for many scientists, objectivity is threatened. The scientific community developed several mechanisms for ensuring the objectivity of its publications. One such mechanism, the peer review system, at least subjects potential articles to the scrutiny of colleagues. The peer review system attempts to guarantee that distributed results uphold standards - namely, that they can be reproduced, and that they can be measured.

Though admittedly not perfect, for a while, the peer review system has improved the academic validity of scientific writing - methods, data, and their resulting publications. In the ultra-competitive scientific world, many scientists relish the opportunity to nit-pick the methods and assumptions of their contemporaries. During my own three summers in research settings, weekly lab meeting banter was always filled with hoots and hollers poking fun at an article received for peer review. This nit-picking, though on the surface driven by competition, serves a valuable purpose in evaluating the methods and conclusions of potential publications.

Some, however, criticize the peer review system for its lack of objectivity, pointing out that scientists evaluating the work of other scientists, work they may have been directly competing against, lacks perspective. Others attack the peer review system's lack of compensation for reviewers, arguing that unless busy scientists are compensated for their thorough efforts, there is no incentive to carefully consider a paper when reviewing. Despite these flaws, however, the peer review system certainly has helped to maintain high standards in scientific journal publications.

Unfortunately, there are no "peers" to review image submissions; only a trained eye would know what signs of image doctoring for which to scan. Major scientific journals - Cell, Nature,Science - utilize images to support and clarify the presentation of data. Today, science is facing an objectivity crisis. Images - pictures, Ultra Violet photographs of DNA gels, films of protein membranes - are undergoing subjective editing, in ways that are parallel to the ways in which words have been manipulated to influence interpretation of data.

Dr. Michael Rossner of Rockefeller University, executive editor of The Journal of Cell Biology, first realized that submitted images had been doctored when the journal began to require digital submissions of images for publication. Unlike hard copies of images, digital images can be magnified with the click of a button, and any area's color compositions analyzed to yield the statistical likelihood that the combination of color pixels appeared naturally. Since this policy change in 2002, over a quarter of all submitted digital images have failed to comply with the journal's image submission guidelines. Rossner attributes the recent increase in image doctoring to the widespread availability of digital media devices, which has removed the technological barriers that previously kept images safe from human alteration or intervention.

The prestigious journal Nature released a widely read article donning a catchy pop-culture reference, "CSI: Cell Biology," attempting to explain this recent uproar surrounding scientific image fabrication and its detection. The piece explains that "[m]ost alterations are harmless: researchers legitimately crop a picture or enhance a faint, fluorescently tagged protein," but that sometimes these innocent alterations "erase valuable data or raise suspicions of fabrication."

As the article points out, scientists doctor images for a wide variety of reasons. Richard Sever, executive editor of the Journal of Cell Science in Cambridge, England explains that the majority of doctoring offenses are "junior people tidying up the image and not realizing that what they're doing is wrong." A few authors, however, have been prosecuted for combining images of cells from several cultures and then assembling the images so they appeared as if all cells were growing in one plate. Sever acknowledges several difficult questions relating to image doctoring in scientific publication. First, many argue that research ethics and morals is an under-represented sector of science education. Perhaps scientists are simply unaware of the potential consequences of their actions. Secondly, in a field where doctoring offenses are performed by both innocent and malicious parties, how is it possible to differentiate between the two? Since some forms of image doctoring are considered tolerable and even necessary and others as scientific fraud, the last and most controversial question asks where the line of acceptability be drawn.

The variety of offenses by scientists submitting research data filled a complete spectrum from innocent and permissible changes, to malicious attempts to fabricate data. For example, most scientific journals permit photo editing, such as changes to an image's brightness, or simple size crops. Some authors of papers, however, would clean up the background of a DNA gel band (a test separating DNA fragments by length) with Photoshop's clone or rubber stamp tool for simply cosmetic purposes. Some would enhance the presence of a band through contrast manipulation, which, though innocent in appearance, can erase valuable data. Others, however, use these same tools to create entire new bands.

Regardless of intentions, image doctoring is a malignant tumor in the scientific community; the honor code has failed under the pressures of "publish, or perish," the post-doc's devil. The United States Office of Research Integrity tries to uphold standards of honesty and reliability in biomedical research. In 1990, only two and one-half percent of all the office's allegations involved the doctoring of images in scientific papers; by 2001, the percentage had leaped up to twenty-six percent. A new safeguard is needed.

The ever-increasing concerns over scientific image integrity have catalyzed a new field: scientific image forensics. Experts use high-resolution enlargements of images, and mathematical algorithms to detail images for signs of doctoring. The most common signs are areas of similar or identical color tones that, given their size, have a low statistical probability of occurring naturally. Often when scientists attempt to remove background noise on a DNA gel, or eliminate fluorescent protein tags, they do so by borrowing a piece of nearby background, which produces these large areas with identical color toning. Other signs are edges and boundaries that are either intentionally blurred, or incredibly sharp and un-realistic. Dr. Farid received a grant from the Federal Bureau of Investigation to assist his research on verifying the authenticity of digital images. Though many have attempted to develop screening processes, Dr. Farid approaches the quandary as simply a data-sorting and analysis problem, explaining that "[a]t the end of the day you need math" to determine beyond a reasonable doubt that the image has been doctored.

Ultimately, the responsibility for verifying the authenticity of scientific images falls to the journals. Because many scientists have failed to remain objective when submitting images to support their research data, it appears that closer scrutiny is necessary. No journal, however, is anxious to ban image manipulation outright. "CSI: Cell Biology" explains that in many experiments, "researchers often have to adjust the relative intensities of red, green and blue fluorescent markers in order to show all three in a single image," which is considered an acceptable form of image manipulation. Dr. Rossner has found the most widely acceptable compromise to date, publishing explicit guidelines for theJournal of Cell Biology, which, in essence, require that any image doctoring must not be part-specific. In other words, lightening or darkening an image is acceptable so long as the alteration affects the entire image, therefore maintaining the original comparative ratios between areas of the image. Katrina Kelner, a deputy editor of Science, commented of Rossner's guidelines that "[s]omething like this is probably inevitable for most journals." The Journal of Cell Science anticipates releasing image manipulation guidelines within the next three months, and Nature Cell Biology now requires submission of the original digital file alongside any image submitted. The only other doctoring guideline widely supported by editors (but not yet in place by a major international journal) is for authors to include a list of image adjustments made to any submitted image.

Journal editors impose image alteration guidelines reluctantly. As a whole, editors lament their necessity, but feel strongly that the alternative - unregulated publication of images alongside articles - could prove detrimental and destructive to the scientific community.

In addition to the cost of the fraudulent research itself, image doctoring as a method of falsifying data costs the scientific research industry billions of dollars. Research within the scientific community is built cumulatively, scientists assuming that journal-published results from one research group are repeatable and therefore are a suitable platform from which to begin their own research. These assumptions save the science community valuable time and money, enabling researchers to move forward in designing investigations, instead of repeating already-proven results. A false platform or foundation, however, might waste years of a scientist's career, as well as valuable research dollars, to the detriment of the scientific community as well as the public it serves.

It is imperative that the scientific community as a whole unite to design a system for maintaining the integrity of its publications. Whether oversight of images falls to the journals, or another institution, it is mandatory that scientists reclaim and protect the objectivity and integrity of the information sharing system.