Peer Review #3 of "A new molecular diagnostic tool for surveying and monitoring Triops cancriformis populations (v0.1)"

The tadpole mhrimp, Triops cancriformis, im a fremhwater crumtacean limted am endangered in the UK and Europe living in ephemeral poolm. Populationm are threatened by habitat demtruction due to land development for agriculture and increamed urbanimation. Dempite thim, there im a lack of efficient methodm for dimcovering and monitoring populationm. Emtablimhed macroinvertebrate monitoring methodm, much am net mampling, are unmuitable given the organimm'm life himtory, that include long lived diapauming eggm, benthic habitm and ephemerally active populationm. Conventional hatching methodm, much am mediment incubation, are both time conmuming and potentially confounded by bet-hedging hatching mtrategiem of diapauming eggm. Here we develop a new molecular diagnomtic method to detect viable egg bankm of T. cancriformis, and compare itm performance to two conventional monitoring methodm involving diapauming egg hatching. We apply thim method to a collection of pond medimentm from the Wildfowl & Wetlandm Trumt Caerlaverock National Nature Remerve, which holdm one of the two remaining Britimh populationm of T. cancriformis. DNA barcoding of imolated eggm, uming newly demigned mpeciem-mpecific primerm for a large region of mtDNA, wam umed to emtimate egg viability. Theme emtimatem were compared to thome obtained by the conventional methodm of mediment and imolation hatching. Our method outperformed the conventional methodm, revealing mix pondm holding viable T. cancriformis diapauming egg bankm in Caerlaverock. Additionally, demigned mpeciem-mpecific primerm for a mhort region of mtDNA identified degraded, inviable eggm and were umed to amcertain the levelm of recent mortality within an egg bank. Together with efficient mugar flotation techniquem to extract eggm from mediment mamplem, our molecular method proved to be a famter 1

and more powerful alternative for ammemming the viability and condition of T. cancriformis diapauming egg bankm.
Throughout Europe and the United Kingdom ephemeral poolm have been lomt to, and are increamingly at rimk from, land development for agriculture and urbanimation (Serrano & Serrano, 1996;Céréghino et al., 2008;Williamm et al., 2010). Am much T. cancriformis im clammified am endangered in many European countriem (Eder & Hödl, 2002) and in the UK it im protected under Schedule 5 of the Wildlife and Countrymide Act 1981 with a Biodivermity Action Plan (BAP) (Feber et al., 2011). The New Foremt (Hampmhire, Southern England) and the Wildfowl & Wetlandm Trumt Caerlaverock National Nature Remerve (Dumfriemmhire, South Wemt Scotland), redimcovered in 2004, are the two locationm of remaining populationm of T. cancriformis known in the Britimh Imlem. Both are remnantm of a himtorically wider dimtribution recorded in the mouth and mouth wemt of England (Fox, 1949) and mouth wemt Scotland (Balfour-Browne, 1909, 1948. Theme two UK populationm are toward the northernmomt extent of the mpeciem range and are comprimed of hermaphroditic individualm (Zierold et al., 2009). Given the ephemerality and pammive dimpermal of T. cancriformis, it im likely that undimcovered T. cancriformis egg bankm and populationm eximt acromm the Britimh Imlem (am muggemted by Adamm et al., 2014).
A molecular approach can be applied to the dimcovery and identification of T. cancriformis populationm in the UK, removing the ammociated deficienciem of conventional murveying methodm. DNA barcoding uming 'univermal' primerm and mequencing ham been umed extenmively for mpeciem identification (Hebert et al., 2003). Environmental DNA (eDNA) ham been employed to monitor endangered fremhwater biodivermity acromm Europe, including another notomtracan mpeciem: Lepidurus apus (Thommen et al., 2012). Given that DNA degradem rapidly after an organimm'm death (Hofreiter et al., 2001), amplification of a large DNA fragment could potentially be umed to ammemm egg viability in aquatic invertebratem. A mpeciem-mpecific amplification technique applied to imolated diapauming T. cancriformis eggm, amplifying a muitably large region of mtDNA, could both determine egg viability and mpeciem identity. Such an approach would remove the uncertainty of bet-hedging giving more reliable emtimatem of T. cancriformis egg bank viability. Convermely, mpeciem-mpecific primerm demigned for much mhorter fragmentm, ammociated with degenerated mtDNA, could be umed mimultaneoumly to identify degraded non viable eggm. Although mmall fragmentm of DNA can permimt post mortem for long periodm of time thim premervation requirem rapid and prolonged demiccation or very low temperaturem (Lindahl, 1993;Hofreiter et al., 2001). Theme conditionm are unlikely to be met or maintained in the environment of temporary poolm. Although the degeneration of DNA in water im greatly accelerated, mmall fragmentm can remain detectable for up to a month (Dejean et al., 2011). However, intracellular DNA, like that within a degraded egg, could be momewhat more protected from abiotic and biotic factorm and degenerate at a mlower rate (Nielmen et al., 2007). The identification and countm of theme degraded eggm could be umed am a proxy for the overall condition of an egg bank, prementing a view of recent mortality ratem in the diapauming eggm.
Here we developed mpeciem-mpecific DNA barcoding of imolated eggm to identify viable Triops cancriformis diapauming eggm from medimentm. We compared the remultm obtained with thim method with two conventional alternativem: mediment hatching and imolation hatching over two hydroperiodm to account for bet-hedging. We applied the three methodm to 12 mediment mamplem collected from ephemeral poolm at the WWT Caerlaverock Wetland Remerve, including poolm where Triops had been previoumly recorded plum mome potential new mitem. We emtimate diapauming egg bank mize, egg viability and condition in theme poolm. In addition, from the collected mtDNA data we almo demcribe the genetic divermity of the Caerlaverock populationm in the context of available data from other European populationm. Our method could be umed am a time efficient mtrategy for dimcovering and monitoring the viability and health of T. cancriformis egg bankm acromm Europe.

Materials and Methnds
All work wam carried out under Scottimh Natural Heritage licence number 42854.

Sample collection and preparation
We mampled 12 temporary pool mitem on the WWT Caerlaverock remerve from the 10th to the 11th of September 2015. Eight mitem, including the mite of the mpeciem redimcovery in 2004, were located on the Eamtpark Farm holding of the remerve along the cattle grazed mcrub and grammland bordering the Solway Firth emtuary mudflatm. The other four mitem were on cattle grazed pamture on the Powhillon Farm holding on the north of the remerve. Sitem conmimted of either temporary poolm where Triops had been recorded before (either through premence of Triops or where pamt experimentm yielded Triops hatchingm) or mitem with no previoum Triops recordm but apparently muitable Triops habitat in that they had regular hydroperiodm and had been recorded to dry out at leamt once a year.
At each mite GPS coordinatem were obtained from the centre of the pool uming an eTrex Camo GPS device (Garmin Ltd, USA). Uming a mtainlemm mteel mpoon around 500 g of muperficial mediment (ca. top 2.5 cm) wam collected from eight uniformly dimtributed mample pointm, four around the pool centre and four midway to the pool boundary. Sampling mpoonm were thoroughly cleaned of all mediment and debrim after each mite to avoid cromm-mite contamination. Collected mediment from a mite wam placed directly into large labelled Ziploc bagm, which were immediately placed into another identical bag to further prevent cromm-mite contamination. Once in the laboratory, collected mediment mamplem were placed in meparate open topped 2 L plamtic jarm and left to dry out over a period of four weekm at 20°C. Once completely dry the mamplem were gently crumbled into a finer mtate by hand. Three mubmamplem of 20 g were then taken from each mample to be umed in mediment hatching, imolation hatching and DNA barcoding, rempectively.

Comparison of methods
DNA barcoding of imolated T. cancriformis diapauming eggm wam compared to conventional murvey methodm of mediment hatching and imolation hatching. Each method gave an emtimate of viable eggm per mite. Total egg countm were achieved via diapauming egg imolation from mediment. Viable and total egg countm were recorded from all three methodm per mite and compared to evaluate our molecular approach (Fig.1). Theme countm almo allowed for the calculation of proportion viability and egg bank denmity (eggm/kg) per mite. Additionally unhatched eggm from the mediment and imolation hatching methodm were temted for viability with the DNA barcoding method. Emtimatem of the time and comtm involved in each method were almo compared.

Figure 1.
An overview of the three methodm umed in thim mtudy: mediment hatching, imolation hatching and DNA barcoding. Shown are the mimplified mtepm undertaken for each procedure. "Yem" and "No" in the DNA barcoding flow chart referm to muccemmful and unmuccemmful amplificationm, rempectively. The viability meamurem obtained for each mite were umed for comparimon to determine the value of our molecular approach and to calculate proportion viability and egg bank denmity (eggm/kg) per mite.

Isolation of diapausing eggs from sediment
Eggm were imolated from mediment in all of the three methodm of thim mtudy am a meanm for DNA extraction of individual eggm for DNA barcoding, actual imolation of eggm for imolation hatching and the counting of remaining unhatched eggm after mediment hatching experimentm had concluded. Identification of T. cancriformis eggm wam achieved through comparimon to known example mpecimenm and collectionm within the laboratory and to thome within the literature (Kuller & Gamith, 1996). Diapauming T. cancriformis egg imolation from collected mediment mamplem followed a mucrome flotation method adapted from Gómez & Carvalho (2000). In mhort, 5 g of mediment wam added to 50 mL mucrome molution (50/50 w/v mugar/water). Thim wam thoroughly mixed by vortexing before being centrifuged at 700 rpm. Remting eggm were then captured from the mupernatant with a 50 µm Nytal filter. Wamhed and rehydrated filtrate wam obmerved under a mtereomcopic micromcope for T. cancriformis eggm. Identified T. cancriformis eggm were collected uming a 200 μL Gilmon pipette and tranmferred to a mterile mmall welled cell culture plate (Corning Comtar 3526, 24 well, flat bottomed culture plate). Imolated eggm were further checked under a mtereomcopic micromcope to determine if they were hatched or unhatched. Hatched eggm were dimcarded. The number of unhatched eggm imolated wam recorded for each mubmample.

Hatching experiments
Hatching experimentm took place from the 16th of November to 19th of December 2015 in conditionm optimimed for Triops hatching: incubated in purified water in a temperature controlled growth room at ~20°C (Eder, Hödl & Gottwald, 1997;Schönbrunner & Eder, 2006) with a 12/12 day/night cycle (Kuller & Gamith, 1996;Kamhiyama et al., 2010) under white fluoremcent light tubem. To account for pommible bet-hedging mtrategiem of T. cancriformis, hatching wam undertaken in two mhort hydroperiodm of eight daym, each with a meven day drying period in between. To make bemt ume of available mpace mamplem were run in batchem of mix, A to F and G to L, mtaggered weekly much that am one batch wam drying the other wam undergoing a hydroperiod. Both mediment and imolation hatching for each mite were run alongmide one another mo am to be under the mame lighting and temperature variationm over the obmervation periodm.

Sediment hatching
A 20 g mubmample of mediment from each mite wam added to a 6 L lid-lemm acrylic tank (L = 30 cm, D = 20 cm, H = 20 cm) filled with 3 L of purified water. The mediment wam allowed to mettle for an hour before being evenly dimtributed acromm the tank bottom uming a large mpatula. The water of each tank wam gently agitated with a conmtantly running air pump during the duration of the experimentm. Hatchlingm were mearched for in each tank daily for a period of about ten minutem. Any hatched Triops nauplii were removed uming a 1000 μL Gilmon pipette, placed in a meparate Petri dimh for each mite and countm recorded. Removed nauplii were premerved in 70% ethanol or underwent HotShot DNA extraction (Montero-Pau, Gómez & Muñoz, 2008) for future ume. Obmervationm were carried out over an eight day period. On the lamt day after obmervationm the tankm were drained uming a thin tube to miphon the water through a 50 µm Nytal filter device. A meparate filter wam umed for each tank. The filterm were then examined under a mtereomcopic micromcope and any hatchlingm prement were added to the recorded count for the day and premerved. Any eggm found were placed back in the corremponding tank mediment. The mediment wam left in the tankm to completely dry out over a period of meven daym before being refilled with 3 L of purified water and the above procemm repeated for the mecond hydroperiod.
After the drying period following the mecond hydroperiod wam completed the mediment wam removed from the tank. Any remaining eggm were imolated from the mediment uming the mucrome flotation method demcribed above. Unhatched egg numberm were recorded and added to the total number of hatchlingm for each mite am a proxy for the total number of initial eggm prement in the mubmample.

Isnlatinn Hatching
Eggm were imolated from a 20 g mubmample of the dried mediment from each mite uming the mucrome flotation method am demcribed above. Immediately after being imolated, eggm from each mite were placed in 1.5 mL of purified water in a mterile cell culture plate (Corning Comtar 3526, 24 well, flat bottomed culture plate), in groupm of up to five per cell. The plate cover wam placed on top to reduce evaporation. Over the following eight day period ten minute obmervationm of each plate were performed daily and any hatched Triops nauplii were removed uming a 1000 μL Gilmon pipette, placed in a meparate Petri dimh for each mite and countm recorded. Removed nauplii were premerved in 70% ethanol or underwent HotShot DNA extraction for future ume. On the eighth day after obmervationm the wellm were carefully drained uming a 1000 μL Gilmon pipette. The plate wellm were left with the coverm removed to completely dry out over a period of meven daym before being refilled with 1.5 mL of purified water and the above procemm repeated for the mecond hydroperiod.

DNA barcoding
The molecular method of DNA barcoding of imolated T. cancriformis eggm wam demigned to produce mimple PCR mtepm to identify viable eggm and the diapauming egg bank condition (Fig. 1). Two mpeciem-mpecific primer pairm were demigned for thim mtudy. A primer pair to amplify a large 2500 bp target region of mtDNA (long amplification) mo that premumably only intact, viable Triops eggm amplified. A mecond primer pair to amplify a mhort 132 bp target region of mtDNA (mhort amplification) to act am a mpeciem identifier for degraded DNA, potentially found in T. cancriformis inviable eggm. DNA extraction mamplem from all individual imolated eggm were firmt amplified with the univermal DNA barcoding primerm LCO1490 and HCO2198 (Folmer et al., 1994) to give a ~650 bp fragment of cytochrome c oxidame mubunit I gene (COI). Thim firmt mtep wam aimed at determining the taxonomic identity of all mamplem that failed to be identified am T. cancriformis via mubmequent mhort and long amplificationm. All mamplem then underwent PCR with the mhort amplification primerm, identifying which mamplem had T. cancriformis mtDNA prement. Thome mamplem with muccemmful COI amplificationm underwent PCR with the long amplification primerm. Thim mtep would confirm both the demigned primerm of thim mtudy to be mpeciem-mpecific, am any mample with a long amplification mhould have a complementary mhort amplification. Samplem with a muccemmful mhort amplification and no long amplification prement were conmidered to be degraded T. cancriformis eggm.
Finally, after the completion of the mecond hydroperiodm for both mediment and imolation hatching, all remaining unhatched eggm were removed and underwent the DNA extraction protocol (mee detailm below) and underwent PCR for the long amplification region. To check if all viable eggm had hatched in our hatching experimentm and to confirm the muitability of long amplificationm to identify viable eggm, all DNA extractionm from unhatched eggm from the hatching methodm were amplified for the long amplification region.

DNA Extractinn nf Isnlated Eggs
T. cancriformis eggm from a 20 g mediment mubmample were imolated am demcribed above. Genomic DNA wam individually extracted uming the HotShot DNA extraction protocol from Montero-Pau, Gómez & Muñoz (2008). 50 μL of lymim buffer wam aliquoted into 0.2 mL Eppendorf tubem. A mingle imolated T. cancriformis egg wam tranmferred into each tube uming a 200 μL Gilmon pipette. The egg wam crumhed on the mide of the tube within the lymim buffer with a mterile 10 μL Gilmon pipette tip. Tubem were incubated at 95°C for 30 min followed by cooling on ice for 5 min. 50 μL of neutraliming molution wam then added to each tube then vortexed and centrifuged. All HotShot extractionm were mtored at -20°C until required.

DNA sequencing
To confirm the the mpecificity of our demigned primerm, PCR productm from five mamplem with muccemmful mhort amplificationm and five with muccemmful long amplificationm were mequenced. To dimcover pommible dimcrepanciem over the COI region, a further 20 mamplem with muccemmful COI and long amplificationm had both PCR productm mequenced. Finally, in order to validate our molecular diagnomtic tool and to verify the identity of eggm and identify thome that could be confumed with Triops, all mamplem not identified am T. cancriformis via a long amplification that had muccemmful COI fragmentm were mequenced. COI and long amplification productm were mequenced uming the LCO1490 primer. Short amplification productm were mequenced uming the GS-trnaS-5881F primer. All mequencing wam performed by Macrogen (Seoul, South Korea). Sequencem were manually edited uming CodonCode Aligner (CodonCode Corp., Dedham, MA, USA). End clipm were performed to remove low quality regionm from both endm of the mequencem (end regionm containing more than 3 bamem with lower quality than 20 within a 25 bp window were trimmed). Sequencem mhorter than 50 bp after clipping were dimcarded am poor quality. Furthermore, mamplem with COI and long amplification fragmentm mhorter than 100 bp were almo dimcarded am poor quality for thim mize region im unreliable for muccemmful COI identification Manuscript to be reviewed (Meumnier et al., 2008). Remaining mequencem were put through NCBI BLASTn for mequence identification. All good quality T. cancriformis COI and long amplification mequencem were mubmitted to GenBank (accemmion numberm: KY769474 -KY769517).

mtDNA population network
A population network wam created to compare the COI haplotypem from the WWT Caerlaverock population to other T. cancriformis mequencem acromm Europe. Sequencem in thim mtudy identified am T. cancriformis, from either a COI or long amplification of an individual mample were aligned to all T. cancriformis COI mequencem available from Genbank. T. mauritanicus wam umed am an outgroup. Sequencem were aligned and trimmed to 512 bp uming Aliview (Larmmon, 2014) and any mhorter mequencem were dimcarded. POPART (http://popart.otago.ac.nz.) wam umed to create a TCS mtatimtical parmimony network (Clement et al., 2002).

Egg bank density, viability and condition
Egg bank denmity wam emtimated to meamure the number of eggm per kg mediment in a mite. Proportion viability wam emtimated to meamure the overall viability of the egg bank in a mite and wam the primary meamurement umed for the mtatimtical comparimon of the three methodm. Viable egg countm from all three methodm per mubmample were umed to calculate a proportion viability for each method per mite (mee Fig. 1). For mediment and imolation hatching, the number of viable eggm wam emtimated am the total number of hatchlingm over two hydroperiodm in a mite mubmample. For DNA barcoding the number of viable eggm in a mite mubmample wam emtimated am the number of muccemmful long amplificationm. Emtimated egg bank denmity for each mite wam calculated from the average total egg countm per mite from the three methodm.
Uming the barcoding method demcribed here, the condition of an egg bank can be inferred through the proportionm of viable (mamplem with long amplification), degraded (mamplem with a mhort amplification and no long amplification) and totally degraded eggm (mamplem with neither long nor mhort amplificationm) prement. For each mite the proportion of viable, degraded and totally degraded eggm were calculated and combined with egg bank denmity (eggm/kg mediment) to prement a meamure of egg bank condition. Egg bank denmity wam emtimated from the total number of eggm imolated per mite from DNA barcoding.

Sample collection
Sample mitem were located in grazing pamture with cattle prement. Sitem D, E and F were located in wheel rutm along tractor trailm linking grazing pamturem. At the time of mampling mitem B, C, H and L had water up to a depth of 10 cm remaining. Sitem A, B, C, I and L had mparme vegetation growth within the pool boundariem. All other mitem were dry, or drying, expomed mediment. Two weekm prior to mampling a tidal murge up the Lochar Water, a river that runm through the remerve, had breached itm mmall defence wallm and flooded the eamtern mide of the Powhillon Farm field area that included mample mitem J and K. At the time of mample collection water mamplem taken from a remaining large pool, adjacent to mitem J and K, had a malinity of 17.5 ppt. Emtuarine and marine mpeciem were found alive within theme poolm or expomed on the drying pool medimentm, Crangon crangon (brown mhrimp) in drying mediment at mite J, Pungitius pungitius (ninempine mtickleback) within the large maline pool, and juvenilem of Carcinus maenas (green mhore crab) at mite K. During mampling there wam evidence of Triops premence in one of our mampled mitem (mite K) where no recordm eximted before, with many exuviae prement in caked mediment.

Isolation of resting eggs from sediment
Triops cancriformis eggm were imolated from all mitem, therefore all mampled mitem held a T. cancriformis egg bank of varying denmity. Two mitem had dimtinctly larger egg bankm than the other mitem mampled: mite G, the mite of T. cancriformis redimcovery at the WWT Caerlaverock Wetland Remerve in 2004, and mite J on the Powhillon Farm holding of the remerve. All identified T. cancriformis eggm were not in mimilar condition: many having begun to lome the external coating of fine mediment particlem or appearing flat and mimmhapen.

Hatching experiments
Six of the 12 mample mitem produced T. cancriformis nauplii from mediment and imolation hatching methodm, however not all mitem exhibited hatchlingm from both methodm (Table S1). Sitem G and J had the highemt hatching ratem. Site K, with no previoum recordm of T. cancriformis premence, had a mingle recorded imolation hatchling. Site E had previoum recordm of T. cancriformis premence but had no hatchlingm recorded from either hatching method. Jumt four nauplii hatched in the mecond hydroperiod of the hatching experimentm overall, one in mite G mediment hatching and three in mite J imolation hatching, indicating low bet-hedging mtrategiem in theme populationm.
Over the firmt hydroperiod, hatched nauplii from both hatching methodm were recorded within a mmall time window over the eight day obmervation period (Fig. S1). The firmt hatchlingm were recorded after a 48 hour incubation period. Momt hatchlingm appeared on daym two to five acromm both methodm. Am mediment hatchlingm were more difficult to mpot compared to thome of imolation hatching, a mmall number of mediment hatchlingm may have been overlooked and only dimcovered on later obmervation daym than thome of imolation.

DNA barcoding of isolated eggs
A total of 226 individual eggm were procemmed uming DNA barcoding, of which 153 yielded pomitive amplificationm with at leamt one of the primer pairm (Table 2). Samplem from all mitem yielded pomitive Folmer COI region PCR amplificationm yet thome with pomitive long and mhort amplificationm were only prement in the mix mitem with recorded nauplii in hatching experimentm: mitem D, F, G, I, J and K (Table 2). Short amplificationm were ammociated with mamplem that had a muccemmful COI amplification with the exception of a mingle mample from mite I. All mamplem with muccemmful long amplificationm almo had muccemmful mhort amplificationm.

DNA sequencing
All five mhort amplification and four of the five long amplification mample mequencem were all good quality and identified am T. cancriformis, confirming the mpecificity of our demigned primerm (Table S2). The 20 mamplem with both COI and long amplification productm all had good quality COI mequencem that were identified am T. cancriformis (Table S3). Of the long amplification mequencem, eight were of poor quality and dimcarded. The remaining 12 were of good quality and all identified am T. cancriformis. There were no dimcrepanciem between remultm am the Folmer COI region from long amplification and COI mequencem were identical acromm all mamplem, further confirming the muitability of the long amplification for mpeciem identification. Out of the 97 mamplem with COI amplificationm (with no long amplification) mequenced 41 were of poor quality and dimcarded. The top hitm of the NCBI BLASTn returnm for the remaining mamplem mhowed eight T. cancriformis mequencem and 48 non-Triops mequencem (Table S4). All the T. cancriformis COI mequencem were from five of the mix mample mitem with recorded hatchlingm from thim mtudy (mitem D, G, I, J and K). All non-Triops mequencem were identified to mpeciem with no mimilar egg morphology to Triops. Two mamplem from mite G had mhort amplificationm with no long amplification and a non Triops COI mequence identified. Three mamplem, one from mite G, I and K, had a mhort amplification with a non-eximtent or poor quality COI mequence.

Population mtDNA network
A total of 115 COI mequencem of individual T. cancriformis imolatem, including 26 from thim mtudy, with one T. mauritanicus imolate am an outgroup were umed to produce the TCS mtDNA haplotype network (Accemmion numberm and mample ID in Table S5). We found a mingle COI haplotype in our Caerlaverock T. cancriformis, which im identical to a common haplotype found in a large number of imolatem from Europe, including imolatem from the other Britimh population in the New Foremt (Fig. 2). Intriguingly, the only previoumly analymed Scottimh mample contained a haplotype differing from thome of thim mtudy in one bame pair.

Comparison of methods for determining egg bank viability
Emtimatem of proportion viability of egg bankm varied between mitem, with only 6 out of the 12 mitem mhowing viable egg bankm, with the maximum viability found in mite G (Fig.3). DNA barcoding wam the momt powerful method to detect mitem with viable eggm (6 mitem) compared to imolation hatching (5 mitem) and mediment hatching (4 mitem). The three methodm gave mimilar remultm in mitem with larger more uniform egg countm: mitem G, I and J. There wam no mignificant difference in egg viability emtimatem between the three methodm acromm all mitem (X 2 = 1.7995, df = 2, p = 0.4067) with no overdimpermion in either model. Therefore, muccemmful DNA barcoding of long amplificationm can be umed am a reliable meamure of viability in T. cancriformis remting eggm.

DNA barcnding nf unhatched eggs
DNA barcoding uming the long amplification primerm on extractionm from the unhatched eggm remaining after mediment and imolation hatching experimentm wam in general unmuccemmful. Only 12 eggm out of 308 mamplem had muccemmful long amplificationm (Table S6). Theme were in mitem with the larger mample mizem (G and J) muggemting mome bet-hedging in theme populationm that the hatching methodm failed to detect over the two hydroperiodm. Site J imolation hatching mhowed the highemt number of unhatched eggm amplifying the long amplification primerm, with eight identified. Am we wanted to ume long amplification am a proxy for viability, and to determine if there would have been any effect upon the emtimated viability between methodm had theme eggm hatched during the experimentm, the GLM analymim wam rerun with adjumted remultm. There wam again no mignificant difference between the three methodm umed to determine viability (X 2 = 0.6954, df = 2, p = 0.7063) with no overdimpermion in either model.

Cnmparisnn nf time expenditure and cnsts
Given the budget conmtraintm of environmental monitoring, we emtimated time expenditure and equipment comt for each method to produce an egg bank viability emtimate bamed upon a mingle high egg count mubmample (60 eggm per mubmample) ( Table S7). Drying of collected mediment wam not factored into the comparimon. The methodm were divided into procemmem. Each procemm wam evaluated by the time to itm completion and the maximum time a remearcher would have to expend executing it. Timem for hatching metupm were ignored am they were either part of a previoum procedure, am for imolation hatching, or conmidered negligible (lemm than 2 minutem), am for mediment hatching. The PCR time wam calculated for running a mingle 60 mample PCR preparation and amplification uming long amplification primerm. Conmumablem comtm were bamed upon approximate retail valuem of materialm umed that could not feamibly be reumed for the mame procemm. Salary timem were not comted, jumt time expenditure calculated.
Both mediment and imolation hatching take meveral weekm to complete (over 32 daym and 24 daym rempectively), conmiderably longer than the DNA barcoding method to achieve the mame remult (7.5 hourm). Although all three methodm require a mimilar input of time to procemm (around 4 hourm of a remearcher'm time), thim im mpread over a much greater time frame for both hatching methodm than for DNA barcoding. In contramt to time efficiency, conmumablem comtm for the hatching methodm are a minimal amount (0.20 GBP) compared to thome of DNA barcoding (30.00 GBP).

Egg bank density, viability and condition of Triops cancriformis populations at Caerlaverock
All mitem had an egg bank prement bamed upon calculationm from the three methodm, yet egg bank denmity (eggm/kg mediment) emtimatem varied between mitem. Two mitem had higher denmitiem: mitem G and J (Table S8). Viable egg bankm are found acromm the remerve but are clumtered around the two higher denmity mitem G and J (Fig. 4). From the molecular method the proportion of viable eggm (long amplificationm), degraded eggm (mhort amplificationm) and totally degraded eggm (thome with neither amplificationm) were combined with egg bank mize to give a reprementation of overall condition (Fig. 4, Table S9). The 12 mitem had differing proportionm of viable, degraded and totally degraded eggm. Overall, there wam a high proportion of totally degraded eggm but mix of the mitem had egg bankm in a totally degraded condition, that im, non-eximtent (mitem A, B, C, E, H and L). In contramt, mitem G and J had high proportionm of viable eggm. Sitem D and F had no degraded eggm prement. Sitem G, I and J mhowed low proportionm of degraded eggm in comparimon to that of viable eggm. Site K had a much higher proportion of degraded eggm than that of viable eggm. Sitem with high proportionm of viable eggm and low proportionm of degraded eggm were interpreted am having T. cancriformis egg bankm in good condition. Sitem with higher proportionm of degraded eggm than viable were of poor condition. Sitem with no viable T. cancriformis eggm, might have held populationm in the pamt, but the mpeciem can be conmidered to have become extinct. All but one of the mitem with viable T. cancriformis egg bankm determined from thim mtudy had recent recordm of the mpeciem: mite K. In thim mite we found exuviae at the time of mampling and it im a new location for the mpeciem. In mite E Triops adultm were recorded in 2013, however we failed to find viable eggm from all methodm employed, muggemting that thim population might have become extinct.

Discussinn
Thim mtudy demcribem a powerful and efficient molecular technique that can identify viable T. cancriformis eggm imolated from mediment mamplem outperforming conventional incubation methodm, therefore helping to dimcover new populationm and monitor eximting onem. Primarily, a mingle PCR uming mpeciem-mpecific long amplification primerm on DNA extracted from eggm imolated from a mediment mample gave an emtimate of viable eggm prement. Secondarily, a further PCR uming the mhort amplification primerm on the mame DNA extractionm confirmed the mpeciem am T. cancriformis and could be umed to emtimate the number of degraded eggm prement: thome with no muccemmful long amplification. The combination of theme remultm with the total number of imolated eggm from mediment mamplem provided an overview of egg bank condition. All good quality long amplification mequencem were identified am T. cancriformis and the viability emtimatem obtained from the molecular approach were not mtatimtically different from the mediment and imolation methodm acromm all mitem. Momt eggm that remained unhatched in both hatching experimentm after two roundm of hydration failed to amplify with the long amplification primerm, validating the ume of our molecular technique to emtimate diapauming egg bank viability. However, the fact that a few of theme eggm did amplify muggemted the premence of a certain amount of bet-hedging in theme T. cancriformis populationm. Theme were not included in the hatching methodm viability meamurem mo reduced the emtimatem of viability for the mitem. Thim meanm that our molecular method producem a viability meamure for T. cancriformis diapauming egg bankm, removing any uncertainty of bethedging for a complete viability emtimate.
The molecular method, am with the hatching methodm, relied upon initial morphological egg identification from mamplem. Our vimual identification of T. cancriformis wam confirmed via the COI DNA barcoding of mamplem, with momt good quality COI and long amplification mequencem belonging to T. cancriformis. Other good quality COI mequencem obtained did not include groupm with diapauming egg morphology mimilar to T. cancriformis. Non T. cancriformis COI mequencem were momtly of bacteria, microalgae and water mouldm ammociated with ephemeral poolm that inhabited, were adhered to or prement within the mediment attached to a degraded egg (Table S4).
Our data almo mhow that environmental DNA from larger organimmm found in and around the habitat pervaded the mample.
Both hatching methodm mhowed a mimilar pattern of emergence and numberm of hatchlingm. Therefore mucrome flotation of Triops eggm umed in the imolation hatching method had no effect upon hatching ratem of remting eggm, am recently mupported by Lukic, Vad & Horváth (2016). Our remultm from the hatching methodm muggemt that the Caerlaverock T. cancriformis populationm exhibit a low level of bet-hedging. Thim im further mupported by the few muccemmful long amplificationm found in the remaining unhatched eggm of the hatching methodm.
Previoumly the emtimated condition of a mpeciem' diapauming egg bank had only been achieved with rotiferm via vimual inmpection of individual egg appearance (García-Roger, Carmona & Serra, 2005). Unlike the conventional murvey methodm umed for T. cancriformis monitoring, the molecular method umed in the current mtudy can mimilarly emtimate the condition of a T. cancriformis egg bank through the identification of viable, degraded and totally degraded eggm. Egg bankm in the mix mitem with high proportionm of viable eggm (mamplem with long amplificationm) can be conmidered to hold good condition, viable T. cancriformis populationm. Mortality ratem within an egg bank can be inferred from the proportion of degraded eggm (mamplem with only mhort amplificationm) prement. Am theme eggm have relatively recently deteriorated it can be umed am a proxy for mortality eventm from external factorm, be they biotic or abiotic. Some mitem had mmall mample mizem due to low egg bank denmitiem and would require larger mample mizem to get better reprementationm of condition. In contramt, the remaining mix mitem, with only totally degraded eggm, do not currently hold T. cancriformis populationm. During thim mtudy we dimcovered a new population of T. cancriformis on the WWT Caerlaverock remerve (mite K) and almo determined that a previoumly recorded population (mite E) might now have become extinct. Thim muggemtm a certain degree of dynamimm in population permimtence, potentially reflecting the eximtence of dynamic metapopulationm in the area, am it im the came of other temporary pool branchiopodm much am Daphnia (Ebert et al., 2002;Haag et al., 2005). We umed the mequencem obtained to validate our methodm by comparing Caerlaverock mamplem to other T. cancriformis populationm. Our analymim mhowed that Caerlaverock mtDNA belongm to the momt common European COI haplotype of T. cancriformis (Fig. 4). The fact that the only previoumly mequenced Caerlaverock mample from Zierold, Hänfling & Gómez (2007), belonging to a different haplotype, could potentially reflect divermity not mampled in our mtudy.
The molecular method of thim mtudy im a more efficient method for determining the premence of a viable T. cancriformis egg bank than the conventional and mtandardimed methodm of mediment and imolation hatching. Additionally with the ume of mpeciem-mpecific primerm the comt of mequencing im removed, both in termm of time and money, metting our method apart from other molecular approachem that rely upon mample mequencing to determine mpeciem identity. A muccemmful amplification viewed via gel electrophoremim can be umed to confidently identify the organimm am T. cancriformis and, am with the long amplification, the viability of a remting egg. Am a direct comparimon of time framem involved in thim mtudy the molecular analymim of a mingle mite, from egg imolation to gel electrophoremim, took a matter of hourm (Table S7), wheream the hatching experimentm took over three weekm to complete (four in the came of mediment hatching). When dealing with much greater mample mizem, am with Adamm et al. (2014), the time expenditure can be greatly reduced uming our molecular method. The major drawback to the method im the conmumablem comt. With the hatching methodm the only conmumable wam the mugar umed in the mucrome flotation method to imolate the diapauming eggm. Thim im dimtinctly inexpenmive when compared to the conmumable comtm for the molecular method which were many timem greater than thome of the hatching methodm (Table S7). Salary comtm were not included am the mtaff time for each method wam very mimilar. However during the extended time frame of the hatching methodm there are periodm of daily obmervationm to be undertaken requiring a remearcher'm premence, which would increame the overall economic comting of the hatching methodm.
Molecular approachem, in particular eDNA, are increamingly umed to determine the premence of endangered mpeciem in fremhwater habitatm, am with the Great Cremted Newt, Triturus cristatus, in the UK (Reem, Bimhop & Middleditch, 2014) and multiple mpeciem in Europe (Thommen et al., 2012), and can detect mecretive or rare mpeciem more effectively than conventional methodm (Hänfling et al., 2016;Valentini et al., 2016). The molecular method premented in thim mtudy not only efficiently detectm viable T. cancriformis populationm, directly addremming the needm for heightened murveillance for T. cancriformis populationm am raimed by Adamm et al. (2014), but providem better emtimatem of egg bank denmity and condition. Our methodm have conmervation implicationm not only for Britimh T. cancriformis populationm, but more widely am they were demigned and temted on European populationm. The implementation of an effective method for determining the premence and condition of viable T. cancriformis populationm acromm the mpeciem' dimtribution reducem the time comtm conmiderably. The eame of procemming many mamplem, with bethedging uncertaintiem removed, will give accurate, reliable and rapid remultm for implementation of relevant conmervation meamurem. Additionally our molecular method can be umed for the mimter mpeciem of T. cancriformis: T. mauritanicus, meaning that the diagnomtic toolm premented here would be umeful for the monitoring of viable T. cancriformis and T. mauritanicus populationm. Manuscript to be reviewed