As a birder, I never thought I’d find myself running through traffic chasing house sparrows. Crossing New York City’s Lincoln Highway in Times Square to the nearest patch of green, I saw some of these birds foraging in the grass amidst some sunbathers. Inhibitions to the wind, I whipped out my iPhone and made the camera in my new Birdsnap app zoom in as far as it would go. A sparrow alighted nearby. I snapped my first shot.

The app told me to mark the bird’s eye, then its tail, then began scanning the image. A moment passed before two results were revealed.

Possible matches: peregrine falcon and great horned owl.

Umm, no and no.

With the launch of Birdsnap in May there were 161 unique bird e-guides available in the Apple store alone. Impressive arrays of special features, not to mention prerecorded audio data, make these apps exciting and occasionally controversial substitutes for printed field guides. Birdsnap in particular has redefined the limits of what such e-guides can accomplish: It employs facial-recognition technology to identify birds photographed on an iPhone. It’s both a cool and convenient way to capitalize on the vast amounts of visual data we continually collect on our mobile devices. The app’s integration with photo storage also sidesteps the identification problems many beginning birders face, and allows anyone to revisit special moments with birds they’ve enjoyed before.

But this staggering number of bird ID apps on the market is problematic even as it reflects a growing popularity for birding worldwide. The emerging market for e-guides is already flooded and fragmented, and even the best ones are a long way from fostering a sustained interest in birds among largely untapped mobile audiences.

The barriers to adoption are steep. Among the eight most comprehensive birding apps for North America, only Birdsnap and Merlin, which list 500 and 350 species, respectively, are free. The next cheapest apps are National Geographic Birds ($9.99 for 995 bird species) and Audubon Birds Pro ($9.99 for 821 species). Peterson Birds ($14.99, more than 800 species); iBird Pro ($19.99, 940 species); The Sibley eGuide ($19.99, 810 species); and BirdsEye North America (a live feed app that offers real-time sighting information via eBird for $19.99, with plates for 1,045 species) complete the set.

App size, which averages 700 megabytes, is equally problematic: they are too big to store together on a typical smartphone. This prevents most mobile birders from cross-referencing guides, a practice that becomes crucial as identifications become more challenging, with plumage variations and encounters with rare species.

Just like their paper predecessors, e-guides contain myriad imperfections and idiosyncrasies in data, artwork and function. Birdsnap identifies only 500 of the 914 bird species known to occur in North America, but with the best species-per-dollar rate on the market and a groundbreaking visual dimension in its work flow, it was worth testing in the field. I took it for a test drive on the tree-lined streets of Manhattan’s West Village, where the birds are approachable and encouraging little gardens abound.

My first trial with house sparrows had failed, but I wasn’t deterred. Maybe bad lighting was to blame? Granted, my subject was moving a lot, as birds tend to do, and although Birdsnap’s camera quickly readjusted to compensate for necessary changes in focus and exposure, it was still just an iPhone camera. Besides, if the bird turned or put its head down in grass, the app’s eye–tail equation was easily undone.

Or maybe it was a size issue? A larger bird was probing the grass nearby. Dodging around bikinis and beach towels I followed it back and forth across the green until finally it landed just long enough to allow a couple of well-lit bird snaps—cue the drumroll.

Possible matches: elegant trogon.

Disappointing: Couldn’t Birdsnap see that this was a European starling, one of the most ubiquitous birds in North America?

Thinking fast, I set Birdsnap’s location filter to “Here and Now.” European starlings live almost everywhere, but there isn’t a trogon in the U.S. that would be caught—except on historic occasion—outside its established range in the American Southwest. At least with a location filter the app might rule out such outlandish, wishful thinking.

The app still couldn’t make a match. A New York City summer visitor, the northern mockingbird, was singing in a nearby tree, but if this thing couldn’t tell the difference between a sparrow and an owl, it didn’t seem worth disturbing any more sunbathers for a chancy shot of a bird that probably wouldn’t leave its branch to come forage at my feet.

Birds are small, and iPhone cameras can’t zoom very far. Getting a picture of sufficient quality for the app to ID seemed impossible, and I wasn’t about to buy bread crumbs to scatter on the ground. That seemed like cheating.

So I let someone else do it. Crossing back over the avenue and onto Christopher Street I saw a pair of pigeons pecking at a pizza crust on the sidewalk. I slowed my approach and veered off to one side, trying to look casual as I held my phone at arm’s length for the closest possible shot of some street birds and a piece of garbage.

Success! Birdsnap’s photo magic confirmed these were rock pigeons. All it had needed was a bird big enough to fill the iPhone camera frame, that faced sideways with its head and tail in plain view, and that didn’t fly away in the vicinity of human beings.

Even the best birding apps fall short in the actual practice of birding. But most breakthroughs have a rocky start, and the identification methods employed by these e-guides are bound to create new challenges as the industry pushes into a mobile frontier. For now, though, it’s still up to the birders, not the e-guides, to distinguish right from wrong.