We demonstrate by means of simple, noninvasive methods (analysis of satellite images, field observations, and measuring “deer beds” in snow) that domestic cattle (n = 8,510 in 308 pastures) across the globe, and grazing and resting red and roe deer (n = 2,974 at 241 localities), align their body axes in roughly a north–south direction. Direct observations of roe deer revealed that animals orient their heads northward when grazing or resting. Amazingly, this ubiquitous phenomenon does not seem to have been noticed by herdsmen, ranchers, or hunters. Because wind and light conditions could be excluded as a common denominator determining the body axis orientation, magnetic alignment is the most parsimonious explanation. To test the hypothesis that cattle orient their body axes along the field lines of the Earth's magnetic field, we analyzed the body orientation of cattle from localities with high magnetic declination. Here, magnetic north was a better predictor than geographic north. This study reveals the magnetic alignment in large mammals based on statistically sufficient sample sizes. Our findings open horizons for the study of magnetoreception in general and are of potential significance for applied ethology (husbandry, animal welfare). They challenge neuroscientists and biophysics to explain the proximate mechanisms.Essentially, they looked at a bunch of Google Earth pictures (no shit) and from this have concluded that large mammals can sense where magnetic north is. No mention is made of the fact that the photos would just about all be taken around mid-day. A number of other potential flaws are pointed out in the comments section of Nature News (which is down atm, so I can't link to it). PNAS is a relatively high impact factor journal - it's reasonably exasperating to see this sort of thing published.
While this paper is meant to be serious, there are a number of interesting papers that take the piss out of the field of Geological Information Systems (GIS) written under the pseudonym "Oleg McNoleg", which are well worth tracking down. I particularly like "The integration of GIS, remote sensing, expert systems and adaptive co-kriging for environmental habitat modeling of the Highland Haggis using object-oriented, fuzzy logic and neural-ntwork techniques". It can be found here, and contains some excellent, if very, very nerdy jokes. It's worth noting that a number of the Oleg McNoleg papers have been published in actual journals, proving that even if GIS is a load of utter rubbish, at least those who use it have a sense of humour.
