Carbon dioxide in the Terrarium
Thomas Falk, Swedish Dendrobates Soc.
This article first appeared in Tilgiftaren' the newsletter of the Swedish Dendrobates Society.  We wish to thank Lars Österdahl for translating it from Swedish.

Preparing for a short field trip I wanted to make sure that my four newly metamorphosed E. tricolor would have constant access to food during my absence.  It seemed a good idea to move them to an old 30 litre aquarium that contained a layer of old oak and beech leaves in varying degrees of decomposition that was being used as culture of soil invertebrates like springtails and small woodlice.  The resident fauna was supplemented with a fair number of vitamin dusted small fruit flies.
 Fig 1 Aquarium type

Returning from the trip I found a froglet sitting on the glass wall, and as there seemed to be a lot of fruit flies left, I did not rush to return the frogs to their normal terrarium.  To my dismay, when I decided to do so a few days later, I found the second froglet on the soil surface in a cramped position with its hind legs stretched and the front part of the body raised on the front legs.  A little deeper in the leaf layer a dead froglet was found in about the same position, hind legs stretched and head bent upwards.  The fourth frog was nowhere to be found and it had probably entered the food chain of the soil fauna.
Fig 2 Terrarium type

The two remaining young were quickly moved to a 'hospital tank' with wet moss.  The larger of the two, the one I had found sitting on the glass, quickly recovered and ate flies the same day.  The smaller regained some mobility in its hind legs in the next 24 hours, but it had recurrences of jerks of the hind legs, and I never saw it eat.  It died after another two days.  The surviving frog was kept under observation foursome weeks until it was returned to the original terrarium, where it developed normally.
 Fig 3 Drainage of heavy gas

Had the frogs had been struck by some kind of infection, or could it be that the frogs had become victims of some poisoning?  I realised that I had not modified the aquarium in the way my frog terraria were (fig 2).  This one had no vent openings close to the bottom.  I began to suspect that the frogs had become victims of carbon dioxide poisoning.

Carbon dioxide a threat
Carbon dioxide is an odourless gas that normally makes up 0.04% vol. of the atmosphere.  In mammals expired air normally contains 4% carbon dioxide because of metabolism.  An increased level of carbon dioxide in the air makes it more difficult for the body to get rid of this waste product, and the body reacts with increased breathing activity.  If the carbon dioxide cannot be removed from the blood a lower pH results, which seriously affects the salt balance, a state called "acidosis".  Breathing in a carbon dioxide level of about 5% leads to unconsciousness in a human and higher levels can be lethal.
 Fig 4 Cramped frog

The sensitivity to levels of the gas varies between organisms.  Animals living in the soil or among fermenting material have the highest tolerances.  One soil living species of springtail can survive in up to 35% whereas another, living in the vegetation does not tolerate more than 2%.  It results from the decomposition of leaf litter, from fermentation, and from respiration.

Carbon dioxide is over 60% heavier than air.  In a vivarium with insignificant air movement carbon dioxide will accumulate in the lower parts and a gradient results with the higher concentration at the bottom.  If the terrarium has no opening at a low level a death trap may result.  Other possible waste gases in a terrarium, like methane or ammonia are much lighter than air, and thus easier to vent away.

Carbon dioxide Experiment
In order to test the carbon dioxide hypothesis I got three one-week-old E. tricolor froglets from a friend.  They were placed in a small bottle with an inlet hole under the bottom material (Leca; peat moss; leaves).  A fine meshed screen covered the upper opening.  The frogs soon found their favourite places where they sat after feasting on the fruitflies that had also been introduced.
 Fig 5 Asphyxiated frog

When I saw that everything looked normal I let carbon dioxide enter the bottle through the inlet.  After a few minutes the frogs began showing signs of restlessness, jumping around in an aimless manner.  After another minute the first signs of cramps began to show, and two of the small frogs fell on their backs.  Soon all three lay motionless with their hind legs outstretched and no visible breathing movements under their chins (see fig 4).  To prevent certain death I ended the carbon dioxide influx and ventilated the bottle.  No visible improvement occurred until after about 5 minutes oxygen enriched air was blown into the bottle.  Since no breathing was present, any gaseous exchange had to take place entirely through the skin.

The first signs of life were seen after about 15 minutes and after another 5 minutes all three frogs were sitting upright, breathing normally.  A few hours after the experiment they were all feeding, and with no lasting symptoms.

Experiment with low Oxygen
Since the production of carbon dioxide is often connected with the consumption of oxygen it is possible that the result of this experiment was due to low oxygen rather than high carbon dioxide.  For this reason a second experiment was performed with the same set up and the same frogs.  This time argon gas was used instead of carbon dioxide.  Argon is an inert, non-poisonous and odourless, non-combustible gas.  As the oxygen inside the bottle disappeared, the frogs moved around in a much calmer way than in the carbon dioxide experiment.  After a few minutes they opened their mouths in a strange way, as though yawning.  No cramps were observed, but they sat shrunken (see fig 5).  After another minute the breathing movements ceased and they no longer reacted to touch.  The experiment was terminated, and oxygen-saturated air was blown into the chamber.  After five minutes all three frogs had recovered.

Interpretation
In order to reduce the stress and avoid unnecessary harm to the frogs, the two experiments were performed more rapidly than would be the case under natural conditions.  The experiments show that an increased level of carbon dioxide causes unrest and cramps, whereas the low oxygen situation is characterised by a languid attitude without unrest or cramps.  Therefore it must be assumed that the death of the three original tricolor was not a result of a too low oxygen level.
 Fig 6 Air movement when lid is opened

Simple calculation shows that in a closed compartment the carbon dioxide level can reach lethal levels long before the oxygen level gets too low.  Therefore oxygen depletion is no real risk in a transportation box or a terrarium.

The experiments have strengthened my assumption that the frogs died from carbon dioxide that had accumulated near the bottom of the tank, with a number of other factors playing a part.

1.  The lid lay on glass flanges that were meant to keep springtails etc in.  The only ventilation was a small gauze type net in the cover, which made air movement difficult.
2.  The lighting was a 5W compact fluorescent lamp placed above the cover glass.  This gave minimal convection inside the tank.
3.  Only a small plant was present in the tank and in the lowlight it had little chance of absorbing a significant amount of carbon dioxide.
4.  The leaf litter was "ripe" and the population of soil invertebrates considerable.
5.  Small pieces of banana were fermenting, thus producing carbon dioxide (and alcohol) 
6.  The tank was left almost a week without the usual daily opening of the cover.  This provided ample time for the heavy carbon dioxide to accumulate at the bottom.

How common are carbon dioxide problems?
The loss of my E. tricolor froglets is probably not a unique event in the hobby.  One hears reports of 'unexplained' deaths in cases where old aquaria, plastic containers etc have been used for temporary holding frogs, for instance in quarantine practice or directly after metamorphosis which means when the frogs are extra sensitive to stress.  As long as the cover is lifted every day for the input of fruitflies, the build-up of stagnant carbon dioxide rich air at the bottom is made impossible.  When the terrarium is left unattended for a few days, perhaps with extra fly bottle inside to provide food in the unattended period, the combined effect of the extra fermentation and the poor ventilation can be fatal.  Deaths among a group of froglets or a quarantine tank tend to be blamed on infections or parasites, but the real culprit may well be an unsuitable tank.