Archive | February 2013

The Night Belongs to the Lions

Lions are a different species at night.  During the day, they must remain hidden from their prey.  But there is nothing secretive about a lion’s behavior on a moonless night. There is no skulking, no need to hide.  The lions own the darkness.

The Kalahari Bushmen still live the traditional hunter-gatherer lifestyle of our ancestors, and they frequently hunt at night – but only when the moon is above the horizon and bright enough for human eyes to detect shapes and movements.   Without the light of the moon, say the Bushmen, the night belongs to the lions.  So they divide the night with the lions according to the phase of the moon.

And if there is a lunar eclipse? That is just a hungry lion, placing her paw in front of the full moon, stealing a little extra darkness.

The Serengeti lions don’t feed well on moonlit nights. Lions have fuller bellies around the new moon; they are thinnest at the full moon.  To compensate, they scavenge or hunt more during daylight hours. Wild herbivores are available to lions throughout the night, no matter what the moon is doing. However, humans universally stay up until around 10:00 pm or so and sleep until sunrise.  If we’re going to be out and about at night, we’re out in the evening not before dawn.

By the full moon, the hours after sunset are so bright that you can read a newspaper.  But the next night, the moon doesn’t rise until an hour after sunset; by the fourth night, the darkness persists for over three hours before the moon finally rises.

Outbreaks of man-eating lions have killed hundreds of people in southern Tanzania over the past twenty years.  These are agricultural areas where lions mostly survive on bush pigs – agricultural pests that cause people to sleep in their fields to protect their harvests.  Thus pigs provide the link between lions and vulnerable humans.

Former graduate students Hadas Kushnir and Dennis Ikanda visited survivors and victims’ families and recorded the precise time and date of 450 attacks on humans.  Most occurred between sunset and 10:00 pm, and while the last few nights before the full moon were the safest, the first few nights after the full moon were three-and-a-half times more dangerous.  After enduring the bright evenings prior to the full moon, the lions were hungry, and they mostly attacked people upon the return of evening darkness.

The full moon can feel spooky, and humans have constructed moon-related myths about werewolves and lunacy and Halloween.  But there is no linkage between moon phase and suicide or admission to psychiatric institutions.

So what if the full moon doesn’t make you crazy but just makes you nervous?  What if it keeps you safely indoors for a couple of nights?

The full moon isn’t dangerous in itself, but after a few million years of dividing our nights with lions, it would be surprising if we didn’t somehow sense the monthly dividing line between our time and lion time.

Here you can read the original research article about lions and the full moon.

The Wrong Answers

Ever since I started looking into the results from Season 4, I’ve been interested in those classifications that are wrong. Now, when I say “wrong,” I really mean the classifications that don’t agree with the majority of volunteers’ classifications. And technically, that doesn’t mean that these classifications are wrong in an absolute sense — it’s possible that two people classified something correctly and ten people classified it wrong, but all happened to classify it wrong the same way. This distinction between disagreement with the majority and wrong in an absolute sense is important, and is something I’m continuing to explore.

But for right now, let’s just talk about those classifications that don’t agree with the majority. To first look at these “wrong” classifications, I created what’s called a heat map. (Click to make it bigger.)


This map shows all the classifications made in Season 4 for images with just one species in it. (More details on how it’s made at the end, for those who want to know.) The species across the bottom of the map are the “right” answers for each image, and the species along the left side are all the classifications made. Each square represents the number of votes for the species along the left side in an image where the majority voted for the species across the bottom. Darker squares mean more votes.

So, for example, if you find aardvark on the bottom and look at the squares in the column above it, you’ll see that the darkest square corresponds to where there is also aardvark on the left side. This means that for all images in which the majority votes was for aardvark, the most votes went to aardvark — which isn’t any surprise at all. In fact, it’s the reason we see that strong diagonal line from top left to bottom right. But we can can also see that in these majority-aardvark images, some people voted for aardwolf, bat-eared fox, dik-dik, hare, striped hyena, and reedbuck.

If we look at the heat map for dark squares other than the diagonal ones, we can see which animals are most likely confused. I’ve circled in red some of the confusions that aren’t too surprising: wildebeest vs. buffalo, Grant’s gazelle vs. Thomson’s gazelle, male lion vs. female lion (probably when only the back part of the animal can be seen), topi vs. hartebeest, hartebeest vs. impala and eland(!), and impala vs. Grant’s and Thomson’s gazelle.

In light blue, I’ve also circled a couple other interesting dark spots: other-birds being confused with buffalo and hartebeest? Unlikely. I think what’s going on here is that there is likely a bird riding along with the large mammal. Not enough people classified the bird for the image to make it into my two-species group, and so we’re left with these extra classifications for a second species.

It’s also interesting to look at the white space. If you look at the column above reptiles, you see all white except for where it matches itself on the diagonal. That means that if the image was of a reptile, everyone got it. There was no confusing reptiles for anything else. Part of this is that there are so few reptile images to get wrong. You can see that wildebeest have been misclassified as everything. I think that has more to do with there being over 17,000 wildebeest images to get wrong, rather than wildebeest being particularly difficult to identify.

What interesting things do you see in this heat map?

(Read on for the nitty gritty or stop here if you’ve had enough.)

Read More…

Oh dear God we are going to die, part II

### Finally in Serengeti, but frantically catching up on camera traps and haven’t had internet for *days* – so here’s a reposting from 2010. ####

I have been convinced of this fact many times during my short stay in the Serengeti.  Whether it was upon being startled awake in my tent by the sound of nearby lion roars, or attempting to cross the yawning abyss of the Ngare Nanyuki river in my 1980’s era Landrover, my brain fights a constant turbulent battle against my sympathetic nervous system.  Intellectually I know we are not going to die.  In the Serengeti at least, lions do not break into tents, even though they seem to me kind of like twinkies: a plastic yellow shell with soft human marshmallow stuffing.  And the Ngare Nanyuki, even though I cannot see the ground below me as we drive forward, has been crossed many times before.  Norbert laughs at me sometimes, “Ali,” he says, “Do you really think we are going to die?  To die is hard work.”

Today though, as I sit frozen, staring at the smooth cement in front of the bathroom door, my brain knows that one wrong move, and someone actually could die.  The texts and calls roll in.  “GET OUT. GET OUT of the house!” Writes Laura.  “Close the door with a pole and break the window so it can escape.” Writes Anna.  I talk to Megan on the phone.  “I don’t want to leave,” I say, “because then I don’t know if it has really left.”  She agrees.  It is either a spitting cobra or a black mamba, perhaps one of the deadliest snakes in the world, and it is hiding in our house.

I liked snakes when I was a kid.  I still do, actually.  I think I have my mom to thank for my strange affection towards these scaly, slithering creatures.  Unlike many moms, she had no fear of them, often rescuing them from the middle of the road where they had ill-advisedly decided to sun.  We had a 6-foot long garter snake in our backyard for many years, and tried to catch him and tame him on many occasions with minimal success. Even the rattlesnakes I’ve almost stepped on just curl up into themselves and give a halfhearted warning and watch me leave. I like snakes.

The snake in our house is easily 6 feet long, a deep charcoal gray.  I am convinced it is a black mamba.  I was updating some lion photos on the computer, singing along to Josh Ritter, with my back to front door.  George and Norbert were coming home soon, and leftovers were warming on the stove.  The strange swishing noise took some time to sink in.  It wasn’t a coming car, and it wasn’t the wind.  It wasn’t any part of the normal animal chorus that plays outside our house.  Finally, I stand and turn to investigate and catch the thick gray shimmer of a snake undulating across our cold cement floor.  There is no visceral shudder that shakes me, just the cold, knife-like stabbing fear. If a black mamba bites you, you will be dead in hours.  I am no more than 10 feet away from one of the most dangerous animals that I will ever encounter.  I hold my breath and watch it slither into the bathroom, and then I make the calls to those who have been here for many more years than I.  I close all the other doors in the house, and then climb up onto the table, off the ground, and watch the smooth, empty cement in front of our bathroom door.  I am still waiting.  30 minutes.  45 minutes.  60 minutes.  George and Norbert promise they are coming, that they will bring our next-door neighbor Juma Pili to help.  Yet over an hour later, they still do not show up, do not call. 30 minutes.  45 minutes.  60 minutes.  It is now 3pm.  The malaria retrovirals are making me dizzy and I want to curl up in bed, but I’m not sure where this snake is.  Eventually the men show, armed with kerosene and a long pincher-pole.  They splash kerosene into the crannies of our bathroom, where plumbing worked once, decades ago, where the snake is almost certainly curled up and asleep.  Eventually it will tire of the smell and leave.  So they say.

So life goes back to normal, more or less.  George starts to wash vegetables in the kitchen, I return to staring at the computer screen.  Craig calls to talk about permits.  “Oh, the snake,” he says.  “It’s probably just a spitting cobra – not that poisonous, really.  If you catch it in the face, just wash it out.  You’ll go blind for about 12 hours, but nothing permanent.  Least of your worries.  Now, can you please send the data for…” he goes on to talk about permits and data analysis.  I am only half listening, and with the corner of my eyes I am watching the cold, smooth cement outside our bathroom door, smelling the antiseptic aroma of kerosene.

Least of my worries?  I can think of a million things that I am less worried about than the spitting cobra hiding beneath our bathtub.  But okay, I am not going to die today.  Which is good, because I have way too much work to do.

### Epilogue: Two days later I got a cryptic call from Norbert, our car fundi. At this point my Swahili was rather poor, and his English was marginal. But it turns out it was a spitting cobra. It spat on him (this took a while to decipher), but only on his hands. We called the vets for advice and assistance, and all was, in the end, okay. Though the snake not the least of my worries, and certainly not the least of Norbert’s. ###

What used to be the “lion lab”

## Today’s guest post is from Jessica Timmons, a University of Minnesota undergraduate who has volunteered with the Lion Project since 2010 — before we even dreamed of working with Zooniverse to create Snapshot Serengeti. ###

Before there was Snapshot Serengeti, there was Lion Lab. Lion Lab was located at the University of Minnesota in a small room with two computers and rows upon rows of species reference books, film organized in binders, and beaten, rolled up maps that had seen many days in the field. I liked to think of our “mascot” as a small stuffed lion who I nicknamed Leo that sat on top the main computer’s monitor and watched over those working in lab. Ali’s office was located next door, and many other projects’ researchers had offices in the near vicinity. A bulletin board nearby contained a plethora of bios of the many students who volunteered (just as Snapshot Serengeti volunteers do) to identify species in photographs from the Serengeti.

Lion lab volunteers

Lion lab volunteers

My role as the lead undergraduate researcher and volunteer coordinator consisted of working with the project’s volunteers and researchers, acting as a communication channel so that all knew about the exciting happenings in lab. At first I was in charge of organizing volunteers so that each had ample time to ID; since the project was housed on one computer volunteers had to physically come into lab to work with the data. To foster a sense of community, every couple weeks we would host lab meetings where Ali and Craig would talk about all of their experiences in the field and spark a desire in all of us to want to go the Serengeti, too.

As the project grew, there came a time when it became possible to access the program remotely. This meant that volunteers did not have to come into lab anymore and could identify from anywhere they had internet access. Though we could now work from anywhere with the brand, new Serengeti Live program, I and another dedicated volunteer still came into the lab to identify. We loved the atmosphere and always jammed to the Lion King soundtrack as we worked. It was great to have someone to share exciting photo discoveries with – if one of us would spot a lion in an image we would excitedly tell the other then proceed to examine the photo as thoroughly as humanly possible.

Though my job as volunteer coordinator was now irrelevant, I was still someone volunteers could contact with questions.  Since so many people now had access to our project, Ali decided it would be a great idea to have a core group of the most active volunteers that could brainstorm ways to keep the project moving forward. So it happened that a small group of us would meet once a week to discuss and execute plans to make the identification process even smoother. We made online tutorials, species reference guides, and helped to raise money for the project by sending out rewards to those who supported us through a RocketHub campaign. It was around this time that Ali announced the exciting news that the project would become accessible to all through a partnership with Zooniverse. Snapshot Serengeti was born, and because of the dedicated volunteers and researchers out in the field incredible things are being discovered daily. I feel so lucky that I’ve been able to watch the project grow into something truly extraordinary from its beginnings on one computer in a little lab at the University of Minnesota.