Table 1. Results of a review of the empirical literature on P. aeruginosa siderophore mutants and cheating.

Reference	Producer strain	Non-producer strain	Growth medium	Specific test for cooperation	Location of results in publication	Non-producer start frequency	Cheating observed?	Notes
Griffin et al., 2004	ATCC 15692 (PAO1)	PA06609 (PAO9)	CAA+apotransferrin	Yes	Text	0.5	Variable	Cheating was concluded on the basis of differences in density of pure vs mixed culture, relative fitness not reported.
Harrison et al., 2006	ATCC 15692 (PAO1)	PA06609 (PAO9)	Waxworm	Yes	Figures 3 and 4	0.03–0.9	No	Highest reported cheat relative fitness is 1, at start frequencies⩽0.01
Ross-Gillespie et al., 2007	ATCC 15692 (PAO1)	PA06609 (PAO9)	CAA+apotransferrin	Yes	Figure 3	0.001–0.99	Variable	Cheating observed at starting frequencies⩽0.1; total population density only affected at lowest start frequency
Ross-Gillespie et al., 2007	ATCC 15692 (PAO1)	PAO1 ΔpvdD	CAA+apotransferrin	Yes	Figure 3	0.001–0.99	Variable	Cheating observed at starting frequencies of 0.001, but no effect on total population density. This mutant gains less fitness benefit from co-culture than PAO9 does
Ross-Gillespie et al., 2007	UCBPP-PA14 (clinical)	Spontaneous mutant	CAA+apotransferrin	Yes	Figure 3	0.001–0.99	Variable	Cheating observed at start frequencies⩽0.1, but no effects on total population density. This mutant gains less fitness benefit from co-culture than PAO9 does
Harrison et al., 2008	PAO985	De novo evolution experiment	CAA+apotransferrin/iron,±S. aureus	No	Figures 1 and 2	0
Brockhurst et al., 2008	ATCC 15692 (PAO1)	PA06609 (PAO9)	M9 minimal salts+CAA+apotransferrin	No, but raw data available	Figure 4 & raw data	0.5	Yes	Cheat relative fitness in mixed culture decreases as more resources are supplied to the media (CAA concentration manipulated to alter carbohydrate and amino acid supply)
Kümmerli et al., 2009a	ATCC 15692 (PAO1)	PA06609 (PAO9)	CAA+apotransferrin	Yes	Text	0.33	Yes
Kümmerli et al., 2009b	ATCC 15692 (PAO1)	PAO1 ΔpvdD	CAA+apotransferrin	No
Kümmerli et al., 2009b	ATCC 15692 (PAO1)	PAO1 ΔpvdD/ΔpchEF	CAA+apotransferrin	Yes	Figure 4	0.17–0.83	Yes	Cheat is always fitter than the wild type, even when competed at a starting frequency of 0.83.
Kümmerli et al., 2009c	ATCC 15692 (PAO1)	PA06609 (PAO9)	CAA+apotransferrin	Yes	Figure 2	0.33	Variable	Cheating is time dependent
Harrison and Buckling, 2009	ATCC 15692 (PAO1)	pvdF transposon mutant in MPAO1 background (PA2396-C04::ISlacZ/hah)	CAA+apotransferrin	No, but raw data available	Raw data	0.5	No	Re-analysis of raw data reveals this mutant is less fit then the wild type in mixtures with a starting frequency of 0.5 in planktonic and biofilm culture.
Harrison and Buckling, 2009	ATCC 15692 (PAO1)	Clones evolved from PAO6049	CAA+apotransferrin	No	Raw data	0.5	No	Mutants outcompete the wild type in planktonic mixed culture, but so does their siderophore-pruducing ancestor & they are lab adapted, growing as well as PAO1 in pure culture.
Ross-Gillespie et al., 2009	ATCC 15692 (PAO1)	PA06609 (PAO9)	CAA+apotransferrin	Yes	Figure 2	0.09	No
Ross-Gillespie et al., 2009	ATCC 15692 (PAO1)	PAO1 ΔpvdD	CAA+apotransferrin	Yes	Figure 2	0.09	Variable	Cheating observed at high cell density only.
Ross-Gillespie et al., 2009	UCBPP-PA14 (clinical isolate)	Spontaneous mutant	CAA+apotransferrin	Yes	Figure 2	0.09	Variable	Cheating observed at high cell density only.
Kümmerli et al., 2010	ATCC 15692 (PAO1)	PA06609 (PAO9)	CAA+apotransferrin	Yes	Text	0.5	Yes
Kümmerli and Brown, 2010	ATCC 15692 (PAO1)	PAO1 ΔpvdD	CAA+apotransferrin	Yes	Figure 5	0.5	Yes
Kümmerli and Brown, 2010	PAO6049	PA06609 (PAO9)	CAA+apotransferrin	No
Kümmerli and Brown, 2010	Environmental isolate	Spontaneous mutant	CAA+apotransferrin	Yes	Figure 5	0.5	Yes
Jiricny et al., 2010	ATCC 15692 (PAO1)	PAO1 ΔpvdD	CAA+apotransferrin	Yes	Figures 4 and 5	0.09	Yes
Jiricny et al., 2010	ATCC 15692 (PAO1)	PAO1 ΔpchEF	CAA+apotransferrin	No		0.09
Jiricny et al., 2010	ATCC 15692 (PAO1)	PAO1 ΔpvdD/ΔpchEF	CAA+apotransferrin	No		0.09
Jiricny et al., 2010	11 various isolates	Spontaneous mutants	CAA+apotransferrin	Yes	Figures 4 and 5	0.09	Variable	Cheating observed for 8/11mutants. Three non-cheats (with high pyoverdin production) may be driving the reported trend.
Harrison and Buckling, 2011	PAO6049	Evolved clones	CAA+apotransferrin	Yes	Figure 1	0.05, 0.5	Variable	Cheating observed at low starting frequencies for a minority of mutants (mean relative fitness=1)
Harrison and Buckling, 2011	PAO1 ΔmutS	Evolved clones	CAA+apotransferrin	Yes	Figure 1	0.05, 0.5	Variable	Cheating observed at low starting frequencies only for a majority of mutants, but some can cheat from a starting frequency of 0.5.
Dumas and Kümmerli, 2012	ATCC 15692 (PAO1)	Evolved clones	CAA+apotransferrin	No	Figures 3 and 4	0		Pyoverdine-deficient clones evolved, but their presence did not consistently reduce population growth—some were associated with increased growth.
Dumas et al., 2013	ATCC 15692 (PAO1)	PAO1 ΔpvdD and ΔpvdD/ΔpchEF	CAA+apotransferrin	No	Figure 4	Three-strain mix, each strain at 0.33		Carbon source, pH and temperature determine the relative growth advantage conferred by siderophores in monoculture. In acidic pH, ΔpvdD grows better. Report outcome of simuated competitions based on monoculture growth parameters.
Ghoul et al., 2014	Cystic fibrosis isolate	2x spontaneous mutants	CAA	Yes	Figures 3 and 4	0.1	Yes	Mutant with reduced pyoverdine production cheats on parent strain, and is cheated on by a second spontaneous mutant with even lower pyoverdine production.
Ross-Gillespie et al., 2015	PAO1 ΔpvdD	PAO1 ΔpvdD/ΔpchEF	CAA+apotransferrin	Yes	Figure 2	0.2	Yes	This experiment competed a double pyoverdine/pyochelin knockouts against a single pyoverdine knockout.
Ross-Gillespie et al., 2015	PAO1 ΔpchEF	PAO1 ΔpvdD/ΔpchEF	CAA+apotransferrin	Yes	Figure 2	0.2	Yes	This experiment competed a double pyoverdine/pyochelin knockouts against a single pyochelin knockout.
Kümmerli et al., 2015	ATC 15692 (PAO1)	PAO1 ΔpvdD	CAA+apotransferrin	No, but can be inferred from data supplied	Figures 1,2,3	0.5	Yes	There is a coevolutionary arms race, whereby producers become less exploitable and non-producers become better cheats. Non-producer frequency is negatively correlated with population growth. Non-producers grow less well than producers in monoculture, and contemporary pairs show relative fitness of non-producers is >1 (though graphs suggest this effect is small).
Andersen et al., 2015	Cystic fibrosis isolates	CF isolates	No culture	No	Text			Cheating inferred by sequence of mutations affecting pyoverdine production and uptake. Very few mutations reported in pyoverdine biosynthetic loci and none pyochelin loci. Most mutations are in pvdS.
Ghoul et al., 2016	ATCC 15692 (PA01)	PA06609 (PAO9)	CAA+apotransferrin	Yes	Figures 1a and 2a	0.02–0.10	Variable	Mutant cheats only if added to producer cultures before the onset of stationary phase.
Leinweber et al., 2017	ATCC 15692 (PA01)+eGFP tag	PAO1 ΔpvdD/ΔpchEF+mCherry tag.	CAA±apotransferrin	Yes	Figures 1,2c and 3,Supplementary Figure S1	0.1, 0.5, 0.9	Yes	Mutant was fitter than the wild type in shaken liquid medium (no spatial structure) and as fit as the wild type in static medium or medium solidified with agar (spatial structure present) when apotransferrin was added, regardless of starting frequency. No cheating observed when apotransferrin was not added.
Vasse et al., 2017	ATCC 15692 (PA01)	PAO1 ΔpvdD	CAA+apotransferrin	Yes	Supplementary Figure S1 and Figure 1	0.15, 0.45, 0.75	Yes	When progressively higher concentrations of gentamicin are added to the medium, mutants lose their growth disadvanatge in monoculture but gain a larger benefit from co-culture with the wild type.

Abbreviation: CAA, casamino acids medium.