A culture system using primary cultures of zebrafish male germ cells³, in which the differentiation from spermatogonia to functional sperm can occur in vitro, provides an opportunity to produce GM sperm under entirely in vitro conditions. Recently, we demonstrated the successful integration of a retrovirus in cultured sperm and the successful rapid generation of transgenic zebrafish lines through a simple in vitro fertilization using these infected sperm cultures⁴. The transgenic fish are not mosaic and transmit the provirus to their offspring in a Mendelian fashion. This research represents the first occurrence of the generation of a transgenic animal from GM sperm. On the other hand, this experiment implies that a particular developmental stage of male germ cells is acceptable for a virus gene but others are not. Furthermore, it elicits the question whether germ cells have lower sensitivity for infection from viruses than somatic cells.

The efficiency of the production of transgenic fish was considerably lower than expected when considering the multiplicity of infection (MOI) factor⁴. Despite the addition of retrovirus with a MOI of nearly 10 (that is, the addition of nearly 10 infectious virions per zebrafish PAC2 cell) five transgenics were obtained from 89 fish (5.6 % efficiency) when cultured for 12 days. Because we used enzymatically dissociated cells of adult mature testes and removed the sperm, we were permitted to initiate the culture using germ cells at various stages of differentiation and development ranging from spermatogonia to spermatids. If sperm from germ cells that are not susceptible to virus infection contribute to the fertilization, the efficiency for the production of transgenics decreases. Eight to nine days are required for sperm to differentiate from actively dividing cells, such as a premeiotic spermatocyte, in this culture system³. When the length of the culture period is as short as eight days, the efficiency of transgenic production becomes lower. This implies that later stages of male germ cells, such as postmeiotic spermatocytes and spermatids, may not be sensitive to retroviral infection. Hence, the determination of the developmental stage at which male germ cells are competent to receive a foreign gene is of considerable interest. Enriching spermatogonia by employing transgenic fish expressing GFP using the vas promoter⁵, which expresses GFP in spermatogonia and spermatocytes prior to the first meiotic division, and in vitro differentiation in culture will reveal the answer.

Germ cells appear to have a low sensitivity for infection by the retrovirus, as noted by observing germ cell infection at various stages of development ranging from spermatogonia to spermatids (as described above). Pluripotent stem cells possess mechanisms to prevent viral transduction by expressing specific proteins that bind the murine leukemia viral (MLV) long terminal repeat, thus preventing retroviral gene transcription⁶. Recently, De Miguel and Donovan reported a greater infection efficiency of the avian leucosis viruses to mouse spermatogonia than MLV⁷. Germ cells might possess the same mechanisms, resulting in lower sensitivity for infection from such viruses. Because these cellular mechanisms would be conserved in vertebrates, the zebrafish system will facilitate understanding the susceptibility of male germ cells to infection with other types of viruses and for transfection by non-viral techniques such as electroporation.

Attempts to insert foreign genes into the genome of male germ cells of mammals are mainly performed in the mouse and the rat. The most successful result is obtained when spermatogonial stem cells that have differentiated into functional sperm after transplantation into the testes are infected with viruses and then used to generate transfected sperm and transgenic animals⁸. However, male germ-line stem cells, which are needed for transplantation into the testes, require a long time for differentiation into sperm. Therefore, the successful generation of GM sperm requires an in vitro culture system for male germ cells.

Spermatogonia, which proliferate in the testis by mitosis, are called type A at the earlier stage, including the undifferentiated stem cell, and type B for the later stage of the differentiating state. Spermatogonia type A can be infected with a retrovirus and become immortalized cells by introduction of telomerase gene⁹. Although the cells differentiate into spermatids in response to stem cell factor (or Steel factor), it has not been demonstrated that the spermatid can produce a normal mouse and that their genes can be transmitted into their germ line. Recently, normal and fertile mice were produced by nuclear injection using cultured round spermatids from mouse primary spermatocytes, but the introduction of foreign genes into the spermatocyte has not been reported¹⁰. Therefore, in regards to the applicability of GM sperm to mammals, the question is whether foreign genes can insert into the spermatocyte. If the spermatocyte is competent to accept foreign DNA, we are ready to produce transgenic animals by means of in vitro-cultured spermatids. If not, because the production of GM sperm in zebrafish indicates that retroviruses infect at least spermatogonia type B, the requirement for the mammalian system is to affect the method by which spermatogonia type B differentiate into normal spermatids in vitro. The factors required for the differentiation of spermatogonia type B into spermatids should be conserved in vertebrates, and, therefore, may be determined by the culture system of zebrafish male germ cells.

With this new approach using GM sperm, the transgenic animals are not mosaic, so screening for transgenics is reduced and an entire generation of breeding is skipped, saving significant time and money. This same procedure, as used in the zebrafish system, can be used in fisheries fish. Potentially the most exciting application of generating transgenic sperm from primary cultures is the potential to apply a similar technology to mammals. Using zebrafish cultures as a starting point, many fundamental aspects of culturing spermatogonia and their transfection in vitro can be determined, with the ultimate goal of using the technique for many animal systems.

1. Palmiter RD and Brinster RL. (1986) Ann. Rev. Genet. 20: 465-499.

2. Lavitrano M et al. (2002) Proc. Natl. Acad. Sci. USA 99: 14230-14235.

3. Sakai N. (2002) Development 129: 3359-3365.

4. Kurita K, Burgess SM and Sakai N. (2004) Proc. Natl. Acad. Sci. USA 101: 1263-1267.

5. Krovel AV and Olsen LC. (2002) Mech. Dev. 116: 141-150.

J. Virol. 65: 382-388.

7. De Miguel MP and Donovan PJ. (2003) Biol. Reprod. 68: 860-866.

8. Nagano M et al. (2001) Proc. Natl. Acad. Sci. USA 98: 13090-13095.

9. Feng et al. (2002) Science 297: 392-395.

10. Marh J et al. (2003) Biol. Reprod. 69: 169-176.

If the GM line is sufficiently inferior to wild-type lines in one or more fitness components, the transgene should not spread and there is no risk of harm. However, if the GM line has some advantage compared to the wild-type line in at least one fitness component, the transgene might spread, even if the GM line fares poorer in other fitness components. Two possible ecological harms might result if a transgene spreads into a wild population of conspecifics: an "invasion" harm, if transgene spread transforms the wild population into a genetically modified population which might adversely affect other species in the community, and an "extinction" harm if the invaded wild-type population is eliminated during transgene spread. In the latter case, both wild-type and GM individuals perish. We refer to the latter outcome as the `Trojan gene effect'.

Our model predicts that a Trojan gene effect can occur when GM males have a mating advantage over wild-type males, but GM young have reduced viability relative to wild-type young.

Methodology and Results
Our study organism was the small fish, Japanese medaka (Oryzias latipes). Medaka are of no commercial importance but they possess several attributes that make them excellent model organisms that enabled us to obtain data that are applicable to fish species of commercial interest. Males in our line of GM medaka were 83% larger than wild-type males as a result of the insertion of a salmon growth hormone gene (sGH) driven by a metallothionine (Mt) promoter. The gene construct was inserted into medaka eggs by microinjection just after fertilization.

In a recent study, we found that GM males have a mating advantage over wild-type males⁴. The GM males, by virtue of their larger size, could control access to sexually receptive females better than wild-type males, and GM males were also preferred as mates by females. As a result, the GM males obtained more than 75% of all matings when in competition with wild-type males. In an earlier study, we demonstrated that GM juveniles had only 70% of the viability of wild-type juveniles. Thus, the salmon growth hormone gene with its promoter has opposing effects on two fitness components in our GM medaka: one effect (the male mating advantage) causes the transgene to spread into a population of wild-type individuals; the other effect (the viability disadvantage) leads to the demise of the entire population.

During our mating trials, we also discovered that the competitively disadvantaged wild-type male medaka attempt to counter the mating advantage of GM males by employing alternative mating tactics to sire offspring. The wild-type males use two alternative mating tactics to reproduce: they disrupt matings in progress and attempt to gain access to the previously mating female; and they join GM males while they are mating with females and release their own sperm (Figure 1).

Figure 1. A female medaka (center) is spawning simultaneously with a GM male (on her right) and a sneak wild-type male (on her left). The GM male initiated the mating and the sneak male joined the mating individuals. Parentage analyses indicate that such sneak males fertilize about 23% of the eggs. Photo by Clark Gedney.

In many species, it is common for some types of males to have a mating advantage over other types of males. In many organisms, including many fish, the advantaged males are the larger ones. It is also common in many organisms that the disadvantaged males use a mating tactic different from that used by the advantaged males to obtain mates or fertilize eggs. Usually, the disadvantaged males cheat by using sneak behaviors. Sneak tactics are common in salmonid species as well as other fish species of commercial importance. Thus, when we detected sneak tactics in wild-type medaka males, we were interested in exploring the reproductive consequences of alternative mating tactics. In particular, we wondered if wild-type males sire enough offspring by using a sneak tactic to stop the Trojan gene effect. Such an outcome is possible because fertilization success by sneak males reduce the number of offspring sired by mating GM males, hence dilute the GM mating advantage.

To explore this possibility, we extended our original computer model to include alternative mating tactics. We did this by creating two variables: the chance that matings initiated by GM males would include sneak wild-type males, and the degree of success of the sneak males in fertilizing eggs. Our data on medaka revealed that sneaking wild-type males joined about 43% of the matings initiated by GM males. In addition, parentage analyses that we conducted using a microsatellite locus indicated that these sneak males fertilized about 23% of the young when they joined matings.

Our model predicted that the degree of reproductive success observed in wild-type male medaka when using a sneak tactic was insufficient to halt the Trojan gene effect. Our model predicted that population extinction might be allayed only when sneak wild-type males join at least 50% of the matings of GM males and fertilize at least 40% of the eggs.

Conclusions
As far as we are aware, our study is the first to demonstrate for any species that GM males could obtain a mating advantage over wild-type males. Such a mating advantage is a crucial element for a Trojan gene effect. The other crucial element is a viability disadvantage of GM individuals relative to wild-type individuals, which we also observed in our GM medaka line. Thus, our model predicts that if our line of GM medaka invaded a wild medaka population, a local extinction event should occur. The same would be true of any other species, including commercially important species, if GM males have a mating advantage and the young they produce have a survival disadvantage.

How could a Trojan gene effect be avoided? Several ways are possible but none seem likely. GM lines could be constructed to be infertile, for example by inducing triploidy. However, for this control to be effective in allaying a Trojan gene effect, the sterilization process would have to be 100% effective. This is particularly true for a fish species in commercial production because many thousands of eggs are involved. Alternatively, natural selection might provide a long term solution by favoring GM offspring with higher survivorship, thus reducing or eliminating the GM viability disadvantage. Unfortunately, such selection may be too slow to counter the strong selection produced by a mating advantage. As a result, natural selection might be powerless to halt the Trojan gene. Another possibility is that wild-type males might obtain sufficient reproductive success to retard the spread of a transgene by using an alternative mating tactic. For this possibility, the wild-type males would have to fertilize a significant proportion of the eggs produced by females. Usually such alternative mating tactics involve some form of sneak mating. Sneak tactics result in the phenomenon known as sperm competition; that is, the sperm produced by multiple males vie to fertilize the same set of eggs. Previous studies show that sperm competition can be intense⁵ when multiple males mate with the same female; however, the prevalence of matings involving multiple males in nature is unknown for most species. As predicted by our model, a high incidence of multiple male matings could act to delay a Trojan gene effect, particularly if sneak males are successful in fertilizing a reasonable share of the eggs. This is not to say that GM individuals will be eliminated from a wild-type population, however; it only suggests that population extinction should take much longer or might even be deferred completely. In the latter case, a GMO release could result in an invasion outcome, producing a population with a different behavioral profile and potential for novel ecological effects on other species.

1. Reichhardt T. (2000) Will souped-up salmon sink or swim? Nature 406: 12-14.

2. Muir WM and Howard RD. (1999) Possible ecological risks of transgenic organism release when transgenes affect mating success: sexual selection and the Trojan gene hypothesis. Proc. Nat. Acad. Sci. 96: 13853-13856.

3. Muir WM and Howard RD. (2002) Assessment of possible ecological risks and hazards of transgenic fish with implications for other sexually reproducing organisms. Transgenic Research 11: 101-114.

4. Howard RD et al. (2004) Transgenic male mating advantage provides opportunity for Trojan gene effect in a fish. Proc. Nat. Acad. Sci. 101: 2934-2938.

5. Fu P et al. (2001) Tactic-specific success in sperm competition. Proc. Royal Society London B 268: 1106-1112.

P. aeruginosa use rhamnolipids to precipitate biofilm formation². Biofilms are dense growths of microorganisms that tightly adhere to surfaces. These formations help hold bacteria to surfaces and convert loose aggregates such as soil into a solid protective matrix. They also permit microorganisms to produce a favorable environment within the biofilm.

Rhamnolipids, which serve as naturally occurring surfactants, are sold as organic solvents for dispersing oil spills on human-made surfaces and in soil and water. They are used alone or in combination with microorganisms to remove oil slicks produced by boats and oil spills. Their use in soils requires containment of the contaminated site until the dispersed oil can be safely diluted and allowed to travel through the environment. The product is also effective for industrial purposes. Rhamnolipid applications are not restricted by the U.S. Environmental Protection Agency. The EPA has been encouraging research on rhamnolipids for remediation of hydrocarbons, metals, and pesticides.

GM Regulations Then and Now
The Environmental Risk Management Authority (ERMA) approves all field testing and releases of GM organisms under the authority of the Hazardous Substances and New Organisms (HSNO) Act of 1996³. In 2001 the Royal Commission judged their system to be rigorous and sufficient, albeit bureaucratic and onerous at times. Recent law changes have therefore enhanced or clarified the existing system while making changes to improve workability. Formerly, low-risk modifications, such as of laboratory E. coli, were still subject to a similar approvals process as larger-scale field releases through ERMA; such low-level approvals are now handled appropriately by the local Institutional Biological Safety Committees, saving considerable time and cost. Another legal peculiarity required bacterial fermentations over ten litres to be approved differently from smaller batches; such cases are now judged on risk and containment, not capacity.

More changes plugged gaps discovered in applying the law to the real world. For example, the 1996 law covering general release of a GMO such as a modified crop did not allow any post-release conditions to be imposed whatsoever. There is now a `conditional release' category, allowing ERMA to specify ongoing containment standards such as destruction of GM plant reproductive structures, to impose restrictions on use, and to require post-release inspections. Monitoring is treated as an aspect of biosecurity, and civil penalties can now be imposed for violations of post-release conditions such as containment breaches. Approvals can now reflect the relative risk posed by a given GMO on a case-by-case basis, a disappointment for anti-GM lobbyists who have campaigned for a one-ban-fits-all scenario (see below). The law now also protects company confidentiality to some degree, where in the past any trade secrets disclosed as part of field testing applications would be posted on government Web sites for the public (and the competition).

One amendment was urgent: medicines incorporating or manufactured using genetic modification once required multiple approvals. But governments must be able to act swiftly to protect their people from attacks using biological agents. And New Zealand is economically dependent on agricultural exports and therefore on the health of its national herds. A government cannot be hamstrung by a complex approvals process that could prevent a timely response to a human or animal health emergency. Examples of the devastation that can be caused by such events are easy to find, from BSE and Foot and Mouth Disease in livestock through to anthrax. The New Zealand approvals process has now been streamlined.

A Broader View
The government also used the breathing space of the `moratorium' to devise a Biotechnology Strategy for New Zealand⁴. While this is a political document and therefore well padded with promises to improve education, foster partnerships, and create centers of excellence, it also contained much of consequence. One result was a bill out-lawing human reproductive cloning. Another was changes in funding policy, which mean more money for biotechnology and for studies of its impact. Collaborations between state- and privately-funded science are also encouraged with grants going preferentially to market-led research. Such measures build public and business confidence.

The Patent Act (1953) is at last under review. The patentability of plants and animals (but not plant varieties) remains, but humans, their genetic material, and their medical treatments will become unpatentable. The definition of patentability will be harmonized with the Australian one: the invention must be a "manner of manufacture" that is "novel, involves an inventive step," and "has a credible, specific and substantial utility"⁵. Patents will be refused if commercial exploitation will endanger human, animal, or plant life or health or will seriously prejudice the environment.

New Zealand is a signatory of the International Convention for the Protection of New Varieties of Plants 1991 (UPOV91) and offers Plant Variety Rights as a form of intellectual property⁶. These rights are to be updated in order that rights holders can control "essentially derived" varieties that are simple (obvious) changes to their intellectual property. Farm-saved seed will be exempt, however, from charges of infringement, and "breeders' rights" will provide seed companies with the ability to develop new varieties using protected varieties or their essential derivatives. (The status of GM "essential derivatives" is unclear, but patenting of such inventions would be more appropriate in any case.) New Zealand cannot conform fully to UPOV91 until local issues over the ownership or stewardship of natural resources have been resolved with the indigenous Maori. Such issues are unique features of New Zealand, which have impinged on the GM debate. Another is seen in the amendment of the HSNO Act to allow the executive government to commandeer the GMO approvals process under certain conditions that are not purely economic—the importance of cultural, ethical, and spiritual consequences of release is recognized. All New Zealand laws must be interpreted with reference to the principles of the Treaty of Waitangi, the founding agreement between the Maori peoples and the late-arriving Europeans⁷.

Upping The Anti
Meanwhile, the legislature has not been the only organization working late these last two years. Anti-GM protesters successfully dominated the news as the `period of constraint' ended last October. They used a billboard advertisement featuring a naked, blank-faced woman on all fours who was (a) attached to a milking machine via her four breasts, and (b) tattooed (or branded) on her hip with the letters `GM'. On an earlier occasion the same group had invaded the main chamber of parliament and stripped to their pink underclothes. This triumphant run of excellent publicity management came to an end after a confrontation with Fonterra, New Zealand's largest dairy company and main export earner. The group mistakenly alleged that the company would "genetically modify milk with human genes"⁸ when the license in question was for the use of certain DNA marker types in dairy cattle breeding⁹. The debate was not enriched by the egregious intellectual poverty on display and within months the group had disbanded, their figurehead (a British former pop singer) leaving the country vowing to return.

Since that time, only groups such as Aotearoa Genetic Engineering Action Network and Peoples Moratorium Enforcement Agency have been given airtime after trumpeting the use of non-violent "direct action" against "legitimate targets." In truth there has been little activity over the New Zealand summer; many of the anti-GM groups have restricted themselves to updating their Web sites.

The last three years in New Zealand have demonstrated that laws and society will always struggle to keep apace of technological developments and their consequences. We have also seen that the unique mix of cultures, peoples, and environment in a nation will flavor the coming changes in unpredictable ways. It is certain that the GM debate is not yet over, and it is clear that, with the changes described here and the social and political climate, that New Zealand is nevertheless bracing itself for the biotechnology future.

References

1. Disclaimer: opinions expressed are authors' own and do not necessarily reflect those of any organizations or institutions

TOXIC VERDICTS FOR A TOXIN PATENT
Phillip B. C. Jones

Two February decisions transformed Mycogen's patent on synthetic Bacillus thuringiensis toxin genes. The U.S. Patent and Trademark Office (PTO) eliminated 12 of the patent's 14 claims, and the Court of Appeals for the Federal Circuit severely limited the scope of the remaining two. Who said change is good?

PTO
In 1983, Michael J. Adang and John D. Kemp filed a patent application on "Insect Resistant Plants." They assigned the application to Agrigenetics Research Associates Ltd. and then to Lubrizol Genetics Inc. During the next decade, four patent applications were filed in this lineage, which lost John Kemp as a named inventor but gained three new inventors and Mycogen Plant Science Inc. as assignee. U.S. Patent No. 5,380,831 issued in 1995 with 12 claims to methods of designing synthetic Bacillus thuringiensis toxin genes for expression in plants and two claims to a particular Bt toxin gene disclosed in the patent.

A year later, the PTO declared an interference to determine who had been the first to invent the claimed methods. It was not the PTO's idle curiosity that instigated this proceeding; the United States grants patents to the first party who invents the claimed subject matter. Squaring off against the `831 patent were two patent applications assigned to Monsanto; one named Kenneth A. Barton and Michael J. Miller as inventors, and the other named David A. Fischhoff and Frederick J. Perlak.

After hearing testimony and examining laboratory records and exhibits, the PTO's Board of Patent Appeals and Interferences decided that Fischhoff and Perlak had beat the other parties in inventing the methods described in claims 1 to 12 of the `831 patent. Consequently, those claims are no longer enforceable.

The Board released its 200-page decision on February 2. Several weeks later, the Federal Circuit rendered its opinion on the `831 patent.

DOE!
Mycogen v. Monsanto began in May 1995 when the Environmental Protection Agency decided that Monsanto could commercialize its New Leaf potato, a genetically modified plant that expressed Bacillus thuringiensis CryIIIA delta-endotoxin. Mycogen quickly filed a lawsuit against Monsanto in a California district court alleging, among other things, that Monsanto infringed claims 13 and 14 of the `831 patent. Nine years later, the Federal Circuit resolved this issue in Monsanto's favor. The court based its decision on law that did not exist when Mycogen filed its lawsuit.

The outcome of the case hinged on the doctrine of equivalents (DOE). In the United States, a patentee can allege that a defendant infringed patent claims literally or equivalently. A plaintiff can establish literal infringement by showing that every feature itemized in a patent claim is found in the defendant's product or process. If the defendant has not literally infringed, the plaintiff can argue that the accused product or process should be deemed equivalent to the patented invention. The DOE sanctions this expansion of patent claim scope to prevent a person from dodging literal infringement with trivial modifications of the patented invention.

A patent claim provides public notice about the scope of a claimed invention. Since the DOE creates ambiguity about claim scope, courts have devised rules to limit its application, rules such as prosecution history estoppel. A prosecution history is a public record containing correspondence between a patent applicant and a patent examiner. This record may reveal that the applicant persuaded the examiner to issue a patent by revising claims to narrow their scope. During litigation, the patentee will be forbidden ("estopped") from relying on the DOE to recapture claim scope surrendered during patent prosecution to obtain the patent.

Mycogen's patent application had originally included a broad claim to a "synthetic gene designed to be highly expressed in plants comprising a DNA sequence encoding an insecticidal protein which is functionally equivalent to a native insecticidal protein of Bt." The patent examiner, concerned about the unpredictability of foreign gene expression, had rejected this claim. The examiner insisted that Mycogen's application could only support a claim to the synthetic Bt toxin gene described in the patent disclosure. Mycogen eventually pursued claims focused on this gene and obtained a patent.

Monsanto's synthetic Bt toxin gene has a nucleotide sequence that differs from Mycogen's claimed nucleotide sequence by about 19 percent. Since Monsanto does not literally infringe the patent claim, Mycogen alleged infringement under the DOE. Monsanto countered by arguing that the patent applicant had narrowed the claims for purposes of patentability, and therefore, prosecution history estoppel prohibited Mycogen from asserting the DOE against Monsanto. At this point, things get tricky.

Prosecution history estoppel should prevent Mycogen from using the DOE to expand its claims to encompass all forms of synthetic Bt toxin genes. After all, the examiner had rejected this broad scope. But what about equivalents lying in the territory between all synthetic Bt toxin genes and the particular claimed synthetic Bt toxin gene?

Rules about the effect of prosecution history estoppel changed several times since Mycogen had first filed its lawsuit. In 2002, the U.S. Supreme Court's Festo Corp. v. Shoketsu opinion established new rules about prosecution history estoppel, and a year later, the Federal Circuit published its interpretation of the Festo decision.

In the Mycogen case, prosecution history estoppel created a presumption that Mycogen's claims 13 and 14 are limited to the specified Bt toxin gene. According to the new rules, Mycogen has three options for overcoming this presumption: (1) show that Monsanto's alleged equivalent gene would have been unforeseeable at the time that Mycogen had narrowed its patent claims; (2) show that the reason for narrowing the claims was not directly relevant to Monsanto's alleged equivalent; or (3) show that there was "some other reason" that Mycogen could not have described the alleged equivalent.

Mycogen tried to rebut the presumption by arguing that Monsanto's synthetic gene had been unforeseeable. Noting that Mycogen had tried to claim all functionally equivalent Bt toxin genes, however, the court declared that Mycogen had foreseen the possibility of an assortment of synthetic Bt toxin genes that included the accused equivalent gene. This meant that Mycogen did not rebut the presumption on the basis of unforeseeability and could not assert the DOE to expand claim scope to cover Monsanto's gene.

Anyone who owns rights to a patent on a gene or protein should take heed of this case. The decision does not mention evidence that the patent applicants had foreseen Monsanto's Bt toxin gene. Rather, the patent applicants had tried to use a functional description to claim a genus of Bt toxin genes. Yet the court decided that the patent applicants' wishful thinking had rendered Monsanto's particular gene "foreseeable." The Federal Circuit issued its Mycogen decision as an unpublished opinion that may not be cited as precedent. This does not prevent the court from consistently following the Mycogen v. Monsanto reasoning about unforeseeability in future cases.

Selected References

Norman Ellstrand's new book about hybridization between crops and wild relatives will be of interest to many readers of this newsletter. The book's title is a bit of a mouthful, but its contents are intriguing, informative, and easy to swallow. One of the book's greatest attributes is that it can be understood and appreciated by lay readers and experts alike. With insight, originality, and scholarship, Ellstrand brings together classical and current knowledge about crop genetics, hybridization, and evolutionary ecology in a single, comprehensive treatment. Few academics understand population genetics deeply enough to tackle this job. Even fewer are familiar with empirical studies of both natural and agricultural systems, and none can match Ellstrand's command of how all of these topics apply to in situ germ-plasm conservation, the recent evolution of new weeds, and risk assessment of genetically engineered crops.

The book focuses on the extent to which crops cross-pollinate with wild relatives and the evolutionary consequences of this process. The first section deftly explains basic principles of population genetics. These chapters are filled with interesting examples of how gene flow, selection, and genetic drift shape the genetic structure of plant populations. Hybridization and the long-term persistence of crop alleles in wild populations are explored in detail. The middle part of the book examines evidence for spontaneous hybridization between domesticated plants and their wild relatives. Ellstrand focuses on the 25 most important food crops, including tropical species like cassava, cowpea, coconut, and oil palm. Understanding the genetics and reproductive biology of major food plants is useful for predicting effects of both crop-to-wild and crop-to-crop gene flow.

The third and final part of the book discusses the possible "dangers" of crop-wild mating, with three chapters devoted to genetically engineered crops. Evaluating the consequences of gene flow is much more difficult than showing that it happens. Ellstrand explores these "so what" questions in a very balanced way. To some people, the mere presence of transgenes in other populations represents a type of "genetic pollution." Ellstrand explains why this may be more of a concern to bioethicists and others than to population biologists. His discussion of how crop genes, and transgenes in particular, can have detrimental, neutral, or beneficial effects on genetic diversity is excellent—this section should be required reading for all who wish to understand whether and when specific crops might threaten the biodiversity of local landraces or wild germplasm. He devotes less attention to ecological effects of transgenes, but what is covered is accurate and up-to-date. A final chapter discusses physical and biological options for confining unwanted gene flow.